@pyai/sdk 0.14.0 → 0.16.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +25 -2
- package/dist/cli-omni-template.d.ts +2 -2
- package/dist/cli-omni-template.js +2 -2
- package/dist/dialer/index.d.ts +1 -0
- package/dist/dialer/index.js +1 -0
- package/dist/dialer/managedBrowserAudio.d.ts +21 -0
- package/dist/dialer/managedBrowserAudio.js +69 -0
- package/dist/dialer/managedDialer.d.ts +1 -0
- package/dist/dialer/serverBroker.js +7 -0
- package/dist/index.d.ts +42 -0
- package/dist/index.js +4 -0
- package/package.json +1 -1
- package/src/cli-omni-template.ts +2 -2
- package/src/dialer/index.ts +1 -0
- package/src/dialer/managedBrowserAudio.ts +75 -0
- package/src/dialer/managedDialer.ts +1 -0
- package/src/dialer/serverBroker.ts +6 -0
- package/src/index.ts +33 -0
package/README.md
CHANGED
|
@@ -106,12 +106,12 @@ the instant the socket opens and routes server frames to typed callbacks, so you
|
|
|
106
106
|
is acked but silently dropped, giving you a connected session with zero turns):
|
|
107
107
|
|
|
108
108
|
<!-- omni-starter:start -->
|
|
109
|
-
Requires SDK 0.
|
|
109
|
+
Requires SDK 0.16.0 for `--template omni`. The runnable starter uses ESM,
|
|
110
110
|
Node 20.19+ with an explicit `ws` transport, or Node 22+. MCP 0.5.1
|
|
111
111
|
requires Node 22+. The SDK's REST client continues to support Node 18+.
|
|
112
112
|
|
|
113
113
|
```sh
|
|
114
|
-
npm install @pyai/sdk@0.
|
|
114
|
+
npm install @pyai/sdk@0.16.0
|
|
115
115
|
npx pyai init voice-demo --template omni
|
|
116
116
|
cd voice-demo
|
|
117
117
|
npm install
|
|
@@ -728,3 +728,26 @@ journal. Uploaded audio may remain for up to seven days, job results for up to
|
|
|
728
728
|
30 days, and Recap follows workspace retention. Removing a meeting from the app
|
|
729
729
|
clears cached notes and requests cancellation; it does not erase cloud data or
|
|
730
730
|
the browser's saved audio. See [meeting integration](src/dialer/MEETING_INTELLIGENCE.md).
|
|
731
|
+
|
|
732
|
+
## Managed browser audio
|
|
733
|
+
|
|
734
|
+
SDK 0.15 adds a provider-owned browser audio adapter. Your backend creates the
|
|
735
|
+
PyAI browser session; keep its durable request key and call ID in your existing
|
|
736
|
+
recovery journal. The browser never needs a carrier client constructor or account
|
|
737
|
+
credential.
|
|
738
|
+
|
|
739
|
+
```ts
|
|
740
|
+
import { createManagedBrowserAudio } from "@pyai/sdk/dialer";
|
|
741
|
+
const audio = createManagedBrowserAudio();
|
|
742
|
+
await audio.connect({ access_token, connect_params }, state => updateCallUI(state));
|
|
743
|
+
audio.mute(true);
|
|
744
|
+
audio.sendDigits("123#");
|
|
745
|
+
audio.disconnect();
|
|
746
|
+
```
|
|
747
|
+
|
|
748
|
+
The adapter loads media only when connecting and reports `ringing`, `active` and
|
|
749
|
+
`disconnected`. Disconnecting local audio does not prove the phone call ended:
|
|
750
|
+
request hangup through your server and recover the same call until terminal.
|
|
751
|
+
After an uncertain setup result, recover the original session rather than dialing
|
|
752
|
+
a replacement. An unsupported transport or expanded routing parameters fail
|
|
753
|
+
before any media connection.
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
export declare const omniTemplate: Readonly<{
|
|
2
|
-
"README.md": "# Runnable Omni starter\n\nRequires Node 20.19+ and ESM. Node 20 uses the included `ws` transport;\nNode 22 also works. The separate MCP server requires Node 22+.\nThis server-side example uses a synthetic read-only office-hours tool. Read\n`PYAI.md` when generating the project through the CLI.\n\n<!-- omni-install:start -->\n```sh\nnpm install @pyai/sdk@0.
|
|
3
|
-
"package.json": "{\n \"name\": \"pyai-omni-starter\",\n \"version\": \"0.0.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20.19.0\"\n },\n \"scripts\": {\n \"start\": \"node main.mjs\"\n },\n \"dependencies\": {\n \"@pyai/sdk\": \"0.
|
|
2
|
+
"README.md": "# Runnable Omni starter\n\nRequires Node 20.19+ and ESM. Node 20 uses the included `ws` transport;\nNode 22 also works. The separate MCP server requires Node 22+.\nThis server-side example uses a synthetic read-only office-hours tool. Read\n`PYAI.md` when generating the project through the CLI.\n\n<!-- omni-install:start -->\n```sh\nnpm install @pyai/sdk@0.16.0\nnpx pyai init voice-demo --template omni\ncd voice-demo\nnpm install\n```\n\nInitialization is offline; installing dependencies and running `main.mjs` are\nseparate steps. For an existing checkout of this example, run `npm install`\nin its directory instead.\n<!-- omni-install:end -->\n\nInject `PYAI_API_KEY` through your secret manager or shell environment. Do not\npaste a real key into source, a prompt, or shell history. `.env.example` documents\nthe variable; this program does not load `.env` or saved CLI profiles.\n\nPrepare a 24 kHz PCM16 mono WAV, at most 20 seconds, saying “Please look up the\noffice opening time.” Then run:\n\n```sh\nnode main.mjs caller.wav\n```\n\nRunning the program explicitly opens one bounded Omni session and uploads its\ncaptured reply to Hear. Both consume your key's allowed usage. Automatic retries\nare disabled. No managed number is bought and no phone call is placed.\n\n`session.mjs` waits for configuration and the greeting to drain, sends exactly\none paced input stream (caller PCM or silence), returns the read tool's result,\nand captures output using the rate from `hello.audio_out`. Never run a separate\nsilence timer alongside caller audio. `timing.mjs` is generated from the existing\nOmni evaluation harness's timing utilities.\n\nThe new private `omni-run-*` folder contains `session.wav`, `reply.wav`, and\n`report.json`. The report separates received audio, an office-hours answer\nrecovered by Hear from captured audio, and physical playback (not tested).\nThe office-hours check accepts formatting such as `opens@9a.m`, `9 a.m.` and\n`nine in the morning` in a complete affirmative opening-time sentence. Wrong\ntimes, negation, conflicting or truncated text remain unverified. This is a\nconservative check for the example fact, not a general answer-quality score.\nA synthesis transcript is advisory; it is not evidence that speech arrived.\nThis starter does not negotiate playback receipts. Capture ends after the\nsimulated queue drains and two seconds of quiet, a bounded capture heuristic.\nListen to the saved file when qualifying actual sound quality or completeness.\n\nTo exercise interruption, supply a second WAV saying “Stop speaking now”:\n\n```sh\nnode main.mjs caller.wav interruption.wav\n```\n\nThe second clip starts while reply audio is queued. `onBargeIn` clears that\nsimulated queue, and capture observes at least five seconds after that clip ends.\nAn interrupted answer can correctly fail the full-answer check;\ninspect the separate interruption evidence. This is not a physical speaker test.\nFor a real playback adapter, cancel both queued and currently playing audio.\nReplace the example lookup with authorized application data before product use.\n\nOn a session error, the program exits with JSON containing a recognized machine\n`code` (for example `media_dead` or `invalid_configure`). Unknown server codes\nbecome `unrecognized_server_error`; transport or decoding errors without a\nrecognized transport code become `transport_or_protocol_error`. Raw error text,\ncall IDs and unknown code values are omitted. No automatic retry is attempted;\ninspect the code and your authorized diagnostics before deciding to retry.\n";
|
|
3
|
+
"package.json": "{\n \"name\": \"pyai-omni-starter\",\n \"version\": \"0.0.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20.19.0\"\n },\n \"scripts\": {\n \"start\": \"node main.mjs\"\n },\n \"dependencies\": {\n \"@pyai/sdk\": \"0.16.0\",\n \"ws\": \"^8.18.3\"\n }\n}\n";
|
|
4
4
|
"timing.mjs": "// Monotonic capture timing. Arrival time is distinct from queued playback:\n// a single packet may contain seconds of audio and must finish before settling.\nexport const FRAME_MS = 20;\nexport const REALTIME_GAP_LIMIT_MS = 100;\n// Interruption tests need an observation window after caller transmission,\n// even when an earlier cue/response already finished. These are test bounds,\n// not an inferred engine reply-end or a naturalness threshold.\nexport const INTERRUPTION_OBSERVATION = Object.freeze({\n minimumAfterCallerStreamMs: 5000, quietMs: 2000, maximumAfterCallerStreamMs: 25000,\n});\n\n/** Approximate audible bounds; this is energy detection, not a speech/VAD claim. */\nexport function audibleBounds(pcm, rate, rmsFloor = 160) {\n const window = Math.max(1, Math.round(rate / 100)); // 10 ms windows\n let first = null;\n let last = null;\n for (let off = 0; off < pcm.length; off += window) {\n const end = Math.min(pcm.length, off + window);\n let energy = 0;\n for (let i = off; i < end; i++) energy += pcm[i] * pcm[i];\n if (Math.sqrt(energy / (end - off)) >= rmsFloor) {\n first ??= off;\n last = end;\n }\n }\n return { first, last };\n}\n\nexport function newAudioCapture(started) {\n return {\n started, firstPacketAt: null, lastPacketAt: null,\n firstAudioAt: null, lastAudioAt: null, playbackEndAt: null,\n samples: 0, pcm: [], turnBeginAt: null, eouMs: null,\n latestTurnBeginAt: null, postTurnBeginFirstPacketAt: null,\n postTurnBeginFirstAudioAt: null, lastAssistantTranscriptAt: null, lastCallerTranscriptAt: null,\n callerTranscriptEvents: [], assistantTranscriptEvents: [],\n turnBegins: [], unheardCancellations: [], unheardSupersededTurns: [], tools: [], toolResults: [], kb: null, events: [],\n };\n}\n\nexport function recordTurnBegin(ctx, at, turn, origin = 0) {\n const prior = ctx.turnBegins.at(-1);\n const cancellation = ctx.unheardCancellations.at(-1);\n if (prior && cancellation && !cancellation.admitted\n && cancellation.turn === prior.turn && cancellation.turnBeginIndex === ctx.turnBegins.length - 1\n && Number.isSafeInteger(turn) && turn > prior.turn && at > cancellation.at\n && ctx.samples === 0 && ctx.assistantTranscriptEvents.length === 0\n && ctx.tools.length === 0 && ctx.toolResults.length === 0) {\n cancellation.admitted = true;\n ctx.unheardSupersededTurns.push({ turn: prior.turn, beginAtMs: prior.atMs,\n cancelledAtMs: cancellation.atMs, cancellationEventIndex: cancellation.eventIndex,\n replacementTurn: turn, outputPacketsBeforeReplacement: 0 });\n }\n ctx.turnBeginAt ??= at;\n ctx.latestTurnBeginAt = at;\n // Earlier PCM can be a listening cue. A new reply is still pending even\n // when that cue has already finished playing; retain all of its PCM.\n ctx.postTurnBeginFirstPacketAt = null;\n ctx.postTurnBeginFirstAudioAt = null;\n ctx.turnBegins.push({ atMs: Math.round(at - origin), turn: turn ?? null });\n}\n\nexport function recordUnheardCancellation(ctx, event, at, origin, eventIndex) {\n const prior = ctx.turnBegins.at(-1);\n if (event.event !== \"flush\" || event.reason !== \"turn_merge\"\n || !Number.isSafeInteger(event.cancelled_turn) || event.cancelled_turn <= 0\n || prior?.turn !== event.cancelled_turn || at < ctx.latestTurnBeginAt\n || ctx.samples !== 0 || ctx.assistantTranscriptEvents.length\n || ctx.tools.length || ctx.toolResults.length\n || ctx.unheardCancellations.some(c => c.turnBeginIndex === ctx.turnBegins.length - 1)) return;\n ctx.unheardCancellations.push({ turn: event.cancelled_turn, at,\n atMs: Math.round(at - origin), eventIndex, turnBeginIndex: ctx.turnBegins.length - 1,\n admitted: false });\n}\n\nexport function recordAudio(ctx, pcm, at, rate) {\n if (!(pcm instanceof Int16Array) || !pcm.length) return null;\n const copy = pcm.slice(); // Some transports reuse callback buffers.\n const playbackAt = Math.max(at, ctx.playbackEndAt ?? at);\n const endAt = playbackAt + (copy.length / rate) * 1000;\n const bounds = audibleBounds(copy, rate);\n ctx.firstPacketAt ??= at;\n ctx.lastPacketAt = at;\n ctx.playbackEndAt = endAt;\n const afterTurnBegin = ctx.latestTurnBeginAt != null && at >= ctx.latestTurnBeginAt;\n if (afterTurnBegin) ctx.postTurnBeginFirstPacketAt ??= at;\n if (bounds.first != null) {\n ctx.firstAudioAt ??= playbackAt + (bounds.first / rate) * 1000;\n ctx.lastAudioAt = playbackAt + (bounds.last / rate) * 1000;\n if (afterTurnBegin) ctx.postTurnBeginFirstAudioAt ??= playbackAt + (bounds.first / rate) * 1000;\n }\n ctx.samples += copy.length;\n ctx.pcm.push(copy);\n return { playbackAt, endAt, pcm: copy };\n}\n\n/** Preserve negative gaps: they are overlaps, never missing measurements. */\nexport function captureTiming(ctx, caller, origin = 0) {\n const relative = (at) => at == null ? null : Math.round(at - origin);\n const gap = (at) => at == null || caller.speechOffsetAt == null\n ? null : Math.round(at - caller.speechOffsetAt);\n return {\n ttfbMs: gap(ctx.firstAudioAt),\n anyAudioTtfbMs: gap(ctx.firstAudioAt),\n postTurnBeginTtfbMs: gap(ctx.postTurnBeginFirstAudioAt),\n turnMs: gap(ctx.lastAudioAt),\n packetTtfbMs: gap(ctx.firstPacketAt),\n callerStartMs: relative(caller.startedAt),\n callerSpeechOnsetMs: relative(caller.speechOnsetAt),\n callerSpeechOffsetMs: relative(caller.speechOffsetAt),\n callerStreamEndMs: relative(caller.streamEndAt),\n agentFirstPacketMs: relative(ctx.firstPacketAt),\n agentLastPacketMs: relative(ctx.lastPacketAt),\n agentSpeechOnsetMs: relative(ctx.firstAudioAt),\n agentSpeechOffsetMs: relative(ctx.lastAudioAt),\n agentPlaybackEndMs: relative(ctx.playbackEndAt),\n latestTurnBeginMs: relative(ctx.latestTurnBeginAt),\n postTurnBeginFirstPacketMs: relative(ctx.postTurnBeginFirstPacketAt),\n postTurnBeginSpeechOnsetMs: relative(ctx.postTurnBeginFirstAudioAt),\n lastAssistantTranscriptMs: relative(ctx.lastAssistantTranscriptAt),\n callerMaxFrameGapMs: Math.round(caller.maxFrameGapMs ?? 0),\n method: \"client-monotonic-queued-playback-energy-bounds\",\n };\n}\n\nexport async function streamPcmRealtime(omni, pcm, rate, clock, onFrame = () => {}) {\n const frameSize = Math.max(1, Math.round((rate * FRAME_MS) / 1000));\n const bounds = audibleBounds(pcm, rate);\n const startedAt = clock.now();\n let speechOnsetAt = null;\n let speechOffsetAt = null;\n let previousFrameEnd = startedAt;\n let maxFrameGapMs = 0;\n for (let off = 0; off < pcm.length; off += frameSize) {\n const frame = pcm.subarray(off, Math.min(off + frameSize, pcm.length));\n const at = clock.now();\n maxFrameGapMs = Math.max(maxFrameGapMs, at - previousFrameEnd);\n const durationMs = (frame.length / rate) * 1000;\n omni.sendAudio(frame);\n onFrame(frame, at);\n if (bounds.first != null && bounds.first >= off && bounds.first < off + frame.length) {\n speechOnsetAt = at + ((bounds.first - off) / rate) * 1000;\n }\n if (bounds.last != null && bounds.last > off && bounds.last <= off + frame.length) {\n speechOffsetAt = at + ((bounds.last - off) / rate) * 1000;\n }\n previousFrameEnd = at + durationMs;\n // Pace from the actual send. Never burst late frames to \"catch up\".\n await clock.sleep(Math.max(0, previousFrameEnd - clock.now()));\n }\n return { startedAt, speechOnsetAt, speechOffsetAt,\n streamEndAt: clock.now(), maxFrameGapMs };\n}\n\nexport async function streamSilenceWhile(omni, rate, shouldContinue, clock, onFrame = () => {}) {\n const silence = new Int16Array(Math.max(1, Math.round((rate * FRAME_MS) / 1000)));\n while (shouldContinue()) {\n const at = clock.now();\n omni.sendAudio(silence);\n onFrame(silence, at);\n await clock.sleep(FRAME_MS);\n }\n}\n\nexport async function waitForAgentSettle(getCtx, clock, opts = {}) {\n const started = clock.now();\n const observation = opts.callerStreamEndAt != null;\n if (observation && (!Number.isFinite(opts.callerStreamEndAt)\n || opts.callerStreamEndAt < 0 || opts.callerStreamEndAt > started)) {\n throw new Error(\"Invalid caller observation boundary\");\n }\n const timeoutMs = observation ? INTERRUPTION_OBSERVATION.maximumAfterCallerStreamMs : opts.timeoutMs ?? 25000;\n const settleMs = observation ? INTERRUPTION_OBSERVATION.quietMs : opts.settleMs ?? 2000;\n const deadlineAt = (observation ? opts.callerStreamEndAt : started) + timeoutMs;\n const minimumAt = observation ? opts.callerStreamEndAt + INTERRUPTION_OBSERVATION.minimumAfterCallerStreamMs : -Infinity;\n while (true) {\n const ctx = getCtx();\n const at = clock.now();\n if (opts.isClosed?.()) return { reason: \"closed\", at };\n if (observation && at >= deadlineAt) return { reason: \"timeout\", at };\n const hasResponseAudio = !opts.requireTurnBegin || (ctx.latestTurnBeginAt != null\n && ctx.postTurnBeginFirstAudioAt != null);\n // Synthesis text is progress, not a completion event. It can arrive after\n // early/final audio, so extend the quiet floor without demanding another\n // packet after every advisory. The protocol still has no reply-end marker.\n const activityEnd = Math.max(ctx.playbackEndAt ?? -Infinity,\n ctx.latestTurnBeginAt ?? -Infinity, ctx.lastAssistantTranscriptAt ?? -Infinity,\n observation ? ctx.lastCallerTranscriptAt ?? -Infinity : -Infinity);\n // Do not demand a new post-caller turn_begin: an earlier turn can be\n // coalesced and speak later. This records a bounded quiet observation;\n // downstream response-timing checks can impose a stricter begin boundary.\n if (ctx.firstPacketAt != null && hasResponseAudio && at >= minimumAt && at >= activityEnd + settleMs) {\n return { reason: \"settled\", at };\n }\n if (opts.allowEmpty && ctx.firstPacketAt == null && at >= minimumAt && at - started >= (opts.emptyWaitMs ?? 800)) {\n return { reason: \"empty\", at };\n }\n if (at - started >= timeoutMs) return { reason: \"timeout\", at };\n await clock.sleep(observation ? Math.min(20, deadlineAt - at) : 20);\n }\n}\n\nexport async function withTimeout(promise, ms, message, clock) {\n let timer;\n try {\n return await Promise.race([\n promise,\n new Promise((_, reject) => {\n timer = clock.setTimeout(() => reject(new Error(message)), ms);\n }),\n ]);\n } finally {\n clock.clearTimeout(timer);\n }\n}\n";
|
|
5
5
|
"main.mjs": "import PyAI from \"@pyai/sdk\";\nimport WebSocket from \"ws\"; // Explicit transport makes this work on Node 20 too.\nimport { readFile, writeFile, mkdtemp } from \"node:fs/promises\";\nimport { join } from \"node:path\";\nimport { runSession } from \"./session.mjs\";\nimport { readWav, writeWav, assessAnswer } from \"./audio.mjs\";\n\nasync function main() {\n const [callerPath, interruptionPath, extra] = process.argv.slice(2);\n if (!callerPath || extra) throw new Error(\"Usage: node main.mjs caller.wav [interruption.wav]\");\n if (!process.env.PYAI_API_KEY) throw new Error(\"Inject PYAI_API_KEY through your environment first\");\n const caller = readWav(await readFile(callerPath));\n const interruption = interruptionPath ? readWav(await readFile(interruptionPath)) : undefined;\n const pyai = new PyAI({ apiKey: process.env.PYAI_API_KEY, maxRetries: 0,\n fetch: (url, options) => fetch(url, { ...options, signal: AbortSignal.timeout(30000) }),\n });\n const { report, audio, replyAudio } = await runSession({ pyai, webSocket: WebSocket, caller, interruption });\n const directory = await mkdtemp(join(process.cwd(), \"omni-run-\"));\n const save = (name, data) => writeFile(join(directory, name), data, { mode: 0o600, flag: \"wx\" });\n await save(\"session.wav\", writeWav(audio, report.output_rate ?? 24000));\n const reply = writeWav(replyAudio, report.output_rate ?? 24000);\n await save(\"reply.wav\", reply);\n // Hear the captured bytes, rather than trusting synthesis-advisory text.\n let capturedTranscript = \"\";\n if (report.reply_audio_bytes) {\n try {\n const heard = await pyai.audio.transcriptions.create({ file: new File([reply], \"reply.wav\", { type: \"audio/wav\" }), model: \"pyai-hear\" });\n capturedTranscript = heard.text;\n } catch { report.hear_error = \"Captured audio saved; Hear verification failed\"; }\n }\n const result = { ...report, captured_transcript: capturedTranscript, evidence: assessAnswer(report, capturedTranscript) };\n await save(\"report.json\", JSON.stringify(result, null, 2) + \"\\n\");\n console.log(JSON.stringify({ directory, ...result }, null, 2));\n if (result.evidence.captured_answer !== \"verified_by_hear\" || (interruption && result.evidence.interruption !== \"simulated_queue_cleared\")) process.exitCode = 1;\n}\n\nmain().catch(error => {\n const message = String(error.message).replaceAll(process.env.PYAI_API_KEY || \"\\0\", \"[REDACTED]\");\n console.error(JSON.stringify({ error: message, ...(error.omni_code ? { code: error.omni_code } : {}) })); process.exitCode = 1;\n});\n";
|
|
6
6
|
"session.mjs": "import { newAudioCapture, recordAudio, recordTurnBegin, streamPcmRealtime, INTERRUPTION_OBSERVATION } from \"./timing.mjs\";\n\nexport const realClock = {\n now: () => performance.now(),\n sleep: ms => new Promise(resolve => setTimeout(resolve, ms)),\n};\n\n// Only recognized machine codes may leave the example. Error messages, call\n// IDs and arbitrary code values can contain credentials or customer content.\nconst publicErrorCodes = new Set([\n \"invalid_configure\", \"unsupported_tool_transport\", \"tool_configuration_locked\",\n \"media_dead\", \"unauthorized\", \"insufficient_scope\", \"rate_limit_exceeded\",\n \"concurrency_limit_exceeded\", \"daily_cap_exceeded\", \"credit_exhausted\",\n]);\nconst transportErrorCodes = new Set([\"ECONNRESET\", \"ECONNREFUSED\", \"ETIMEDOUT\", \"ENOTFOUND\", \"EAI_AGAIN\"]);\nexport function omniFailure(error) {\n const code = error instanceof Error\n ? (transportErrorCodes.has(error.code) ? error.code : \"transport_or_protocol_error\")\n : (publicErrorCodes.has(error?.code) ? error.code : \"unrecognized_server_error\");\n return Object.assign(new Error(`Omni failed (${code}); no automatic retry was attempted`), { omni_code: code });\n}\n\n/** One bounded, server-side example. Its playback queue is simulated. */\nexport async function runSession({ pyai, webSocket, caller, interruption, clock = realClock, timeoutMs = 60000 }) {\n const rate = 24000, frameSize = 480, silence = new Int16Array(frameSize);\n if (!(caller instanceof Int16Array) || !caller.length || caller.length > rate * 20) throw new Error(\"Provide 1–20 seconds of caller PCM16 at 24 kHz\");\n if (interruption && (!(interruption instanceof Int16Array) || !interruption.length || interruption.length > rate * 20)) throw new Error(\"Invalid interruption PCM\");\n const started = clock.now(), all = newAudioCapture(started), reply = newAudioCapture(started);\n const report = { input_rate: rate, output_rate: null, playback_sink: \"simulated\", tool_calls: 0, tool_executions: 0, duplicate_tool_calls: 0, flushes_after_interruption: 0, cleared_queue_ms: 0, interruption_requested: !!interruption, caller_started_ms: null, interruption_started_ms: null, advisory_transcripts: [], end_reason: \"timeout\" };\n let session, configuredAt = null, closed = false, failure = null, phase = \"greeting\", offset = 0, queueEnd = started, interruptionSent = false, inputEndAt = started;\n const handled = new Set();\n const fail = message => { failure ??= new Error(message); };\n try {\n session = pyai.omni.connect({\n webSocket, rate,\n configure: {\n voice_id: \"stock_amos_en_us\", language: \"en\", greeting: \"Hello. How can I help?\",\n persona: \"When asked about office hours, call lookup_office_hours. Never guess. After the result, say one short sentence giving its opening_time.\",\n tools: [{ name: \"lookup_office_hours\", description: \"Read the synthetic example office opening time.\", side_effect: \"read\", parameters: { type: \"object\", properties: {}, required: [], additionalProperties: false } }],\n },\n onHello: frame => {\n const match = /^pcm16@(8000|24000)$/.exec(frame.audio_out ?? \"\");\n if (!match) return fail(\"Unsupported Omni output format\");\n report.output_rate = Number(match[1]);\n },\n onConfigured: () => { configuredAt ??= clock.now(); },\n onAudio: chunk => {\n if (!report.output_rate) return fail(\"Audio arrived before its output format\");\n const bytes = chunk instanceof ArrayBuffer ? new Uint8Array(chunk) : new Uint8Array(chunk.buffer, chunk.byteOffset, chunk.byteLength);\n if (bytes.byteLength % 2) return fail(\"Invalid PCM16 output\");\n const view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);\n const pcm = Int16Array.from({ length: bytes.length / 2 }, (_, i) => view.getInt16(i * 2, true));\n const now = clock.now();\n recordAudio(all, pcm, now, report.output_rate);\n queueEnd = Math.max(now, queueEnd) + pcm.length / report.output_rate * 1000;\n if (report.tool_executions) recordAudio(reply, pcm, now, report.output_rate);\n },\n onEvent: frame => {\n if (frame.event === \"turn_begin\") recordTurnBegin(all, clock.now(), frame.turn);\n },\n onTranscript: frame => {\n // An assistant advisory is progress; never treat it as received speech.\n if (report.advisory_transcripts.length < 100) report.advisory_transcripts.push({ role: frame.role, text: frame.text });\n if (frame.role === \"assistant\") all.lastAssistantTranscriptAt = clock.now();\n },\n onToolCall: frame => {\n report.tool_calls++;\n if (handled.has(frame.call_id)) { report.duplicate_tool_calls++; return; }\n handled.add(frame.call_id);\n if (frame.name !== \"lookup_office_hours\") { session.toolResult(frame.call_id, { error: \"Unknown example tool\" }); return; }\n // Replace this read-only fixture with your authorized data lookup.\n report.tool_executions++;\n session.toolResult(frame.call_id, { result: { opening_time: \"9 a.m.\", source: \"synthetic example office\" } });\n },\n onBargeIn: () => {\n const now = clock.now(), queued = Math.max(0, queueEnd - now);\n if (report.interruption_started_ms != null) { report.flushes_after_interruption++; report.cleared_queue_ms += queued; }\n // A real speaker adapter must cancel scheduled and currently playing audio.\n queueEnd = now; all.playbackEndAt = now; reply.playbackEndAt = now;\n },\n onError: error => { failure ??= omniFailure(error); },\n onClose: () => { closed = true; },\n });\n while (clock.now() - started < timeoutMs) {\n if (failure) throw failure;\n if (closed) { report.end_reason = \"closed\"; break; }\n const now = clock.now();\n if (configuredAt == null || !report.output_rate) {\n if (now - started >= 10000) throw new Error(\"Omni connection/configuration timed out\");\n await clock.sleep(20); continue;\n }\n const quietAt = Math.max(queueEnd, all.lastPacketAt ?? now, all.lastAssistantTranscriptAt ?? 0, inputEndAt);\n if (phase === \"greeting\") {\n if (all.firstPacketAt != null && now - configuredAt >= 2000 && now >= quietAt + 800) {\n phase = \"caller\"; report.caller_started_ms = now - started;\n } else if (now - configuredAt > 15000) throw new Error(\"Greeting did not drain within 15 seconds\");\n }\n if (phase === \"reply\" && interruption && !interruptionSent && reply.samples && queueEnd - now >= 200) {\n phase = \"interruption\"; offset = 0; interruptionSent = true; report.interruption_started_ms = now - started;\n }\n const observationEnd = inputEndAt + (interruptionSent ? INTERRUPTION_OBSERVATION.minimumAfterCallerStreamMs : 0);\n if (phase === \"reply\" && reply.samples && now >= quietAt + 2000 && now >= observationEnd && (!interruption || interruptionSent)) {\n report.end_reason = \"quiet_window\"; break;\n }\n // Exactly one producer: each 20 ms slot contains caller PCM OR silence.\n let pcm = silence;\n if (phase === \"caller\" || phase === \"interruption\") {\n const source = phase === \"caller\" ? caller : interruption;\n pcm = source.subarray(offset, offset + frameSize); offset += pcm.length;\n if (offset >= source.length) {\n inputEndAt = now + pcm.length / rate * 1000;\n report[phase === \"caller\" ? \"caller_completed_ms\" : \"interruption_completed_ms\"] = inputEndAt - started;\n phase = \"reply\"; offset = 0;\n }\n }\n await streamPcmRealtime(session, pcm, rate, clock);\n }\n report.audio_bytes = all.samples * 2;\n report.reply_audio_bytes = reply.samples * 2;\n report.elapsed_ms = Math.round(clock.now() - started);\n return { report, audio: all.pcm, replyAudio: reply.pcm };\n } finally { session?.close(); }\n}\n";
|
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
// Generated by scripts/generate-omni-starter.mjs; edit examples/omni-starter instead.
|
|
2
2
|
export const omniTemplate = Object.freeze({
|
|
3
|
-
"README.md": "# Runnable Omni starter\n\nRequires Node 20.19+ and ESM. Node 20 uses the included `ws` transport;\nNode 22 also works. The separate MCP server requires Node 22+.\nThis server-side example uses a synthetic read-only office-hours tool. Read\n`PYAI.md` when generating the project through the CLI.\n\n<!-- omni-install:start -->\n```sh\nnpm install @pyai/sdk@0.
|
|
4
|
-
"package.json": "{\n \"name\": \"pyai-omni-starter\",\n \"version\": \"0.0.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20.19.0\"\n },\n \"scripts\": {\n \"start\": \"node main.mjs\"\n },\n \"dependencies\": {\n \"@pyai/sdk\": \"0.
|
|
3
|
+
"README.md": "# Runnable Omni starter\n\nRequires Node 20.19+ and ESM. Node 20 uses the included `ws` transport;\nNode 22 also works. The separate MCP server requires Node 22+.\nThis server-side example uses a synthetic read-only office-hours tool. Read\n`PYAI.md` when generating the project through the CLI.\n\n<!-- omni-install:start -->\n```sh\nnpm install @pyai/sdk@0.16.0\nnpx pyai init voice-demo --template omni\ncd voice-demo\nnpm install\n```\n\nInitialization is offline; installing dependencies and running `main.mjs` are\nseparate steps. For an existing checkout of this example, run `npm install`\nin its directory instead.\n<!-- omni-install:end -->\n\nInject `PYAI_API_KEY` through your secret manager or shell environment. Do not\npaste a real key into source, a prompt, or shell history. `.env.example` documents\nthe variable; this program does not load `.env` or saved CLI profiles.\n\nPrepare a 24 kHz PCM16 mono WAV, at most 20 seconds, saying “Please look up the\noffice opening time.” Then run:\n\n```sh\nnode main.mjs caller.wav\n```\n\nRunning the program explicitly opens one bounded Omni session and uploads its\ncaptured reply to Hear. Both consume your key's allowed usage. Automatic retries\nare disabled. No managed number is bought and no phone call is placed.\n\n`session.mjs` waits for configuration and the greeting to drain, sends exactly\none paced input stream (caller PCM or silence), returns the read tool's result,\nand captures output using the rate from `hello.audio_out`. Never run a separate\nsilence timer alongside caller audio. `timing.mjs` is generated from the existing\nOmni evaluation harness's timing utilities.\n\nThe new private `omni-run-*` folder contains `session.wav`, `reply.wav`, and\n`report.json`. The report separates received audio, an office-hours answer\nrecovered by Hear from captured audio, and physical playback (not tested).\nThe office-hours check accepts formatting such as `opens@9a.m`, `9 a.m.` and\n`nine in the morning` in a complete affirmative opening-time sentence. Wrong\ntimes, negation, conflicting or truncated text remain unverified. This is a\nconservative check for the example fact, not a general answer-quality score.\nA synthesis transcript is advisory; it is not evidence that speech arrived.\nThis starter does not negotiate playback receipts. Capture ends after the\nsimulated queue drains and two seconds of quiet, a bounded capture heuristic.\nListen to the saved file when qualifying actual sound quality or completeness.\n\nTo exercise interruption, supply a second WAV saying “Stop speaking now”:\n\n```sh\nnode main.mjs caller.wav interruption.wav\n```\n\nThe second clip starts while reply audio is queued. `onBargeIn` clears that\nsimulated queue, and capture observes at least five seconds after that clip ends.\nAn interrupted answer can correctly fail the full-answer check;\ninspect the separate interruption evidence. This is not a physical speaker test.\nFor a real playback adapter, cancel both queued and currently playing audio.\nReplace the example lookup with authorized application data before product use.\n\nOn a session error, the program exits with JSON containing a recognized machine\n`code` (for example `media_dead` or `invalid_configure`). Unknown server codes\nbecome `unrecognized_server_error`; transport or decoding errors without a\nrecognized transport code become `transport_or_protocol_error`. Raw error text,\ncall IDs and unknown code values are omitted. No automatic retry is attempted;\ninspect the code and your authorized diagnostics before deciding to retry.\n",
|
|
4
|
+
"package.json": "{\n \"name\": \"pyai-omni-starter\",\n \"version\": \"0.0.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20.19.0\"\n },\n \"scripts\": {\n \"start\": \"node main.mjs\"\n },\n \"dependencies\": {\n \"@pyai/sdk\": \"0.16.0\",\n \"ws\": \"^8.18.3\"\n }\n}\n",
|
|
5
5
|
"timing.mjs": "// Monotonic capture timing. Arrival time is distinct from queued playback:\n// a single packet may contain seconds of audio and must finish before settling.\nexport const FRAME_MS = 20;\nexport const REALTIME_GAP_LIMIT_MS = 100;\n// Interruption tests need an observation window after caller transmission,\n// even when an earlier cue/response already finished. These are test bounds,\n// not an inferred engine reply-end or a naturalness threshold.\nexport const INTERRUPTION_OBSERVATION = Object.freeze({\n minimumAfterCallerStreamMs: 5000, quietMs: 2000, maximumAfterCallerStreamMs: 25000,\n});\n\n/** Approximate audible bounds; this is energy detection, not a speech/VAD claim. */\nexport function audibleBounds(pcm, rate, rmsFloor = 160) {\n const window = Math.max(1, Math.round(rate / 100)); // 10 ms windows\n let first = null;\n let last = null;\n for (let off = 0; off < pcm.length; off += window) {\n const end = Math.min(pcm.length, off + window);\n let energy = 0;\n for (let i = off; i < end; i++) energy += pcm[i] * pcm[i];\n if (Math.sqrt(energy / (end - off)) >= rmsFloor) {\n first ??= off;\n last = end;\n }\n }\n return { first, last };\n}\n\nexport function newAudioCapture(started) {\n return {\n started, firstPacketAt: null, lastPacketAt: null,\n firstAudioAt: null, lastAudioAt: null, playbackEndAt: null,\n samples: 0, pcm: [], turnBeginAt: null, eouMs: null,\n latestTurnBeginAt: null, postTurnBeginFirstPacketAt: null,\n postTurnBeginFirstAudioAt: null, lastAssistantTranscriptAt: null, lastCallerTranscriptAt: null,\n callerTranscriptEvents: [], assistantTranscriptEvents: [],\n turnBegins: [], unheardCancellations: [], unheardSupersededTurns: [], tools: [], toolResults: [], kb: null, events: [],\n };\n}\n\nexport function recordTurnBegin(ctx, at, turn, origin = 0) {\n const prior = ctx.turnBegins.at(-1);\n const cancellation = ctx.unheardCancellations.at(-1);\n if (prior && cancellation && !cancellation.admitted\n && cancellation.turn === prior.turn && cancellation.turnBeginIndex === ctx.turnBegins.length - 1\n && Number.isSafeInteger(turn) && turn > prior.turn && at > cancellation.at\n && ctx.samples === 0 && ctx.assistantTranscriptEvents.length === 0\n && ctx.tools.length === 0 && ctx.toolResults.length === 0) {\n cancellation.admitted = true;\n ctx.unheardSupersededTurns.push({ turn: prior.turn, beginAtMs: prior.atMs,\n cancelledAtMs: cancellation.atMs, cancellationEventIndex: cancellation.eventIndex,\n replacementTurn: turn, outputPacketsBeforeReplacement: 0 });\n }\n ctx.turnBeginAt ??= at;\n ctx.latestTurnBeginAt = at;\n // Earlier PCM can be a listening cue. A new reply is still pending even\n // when that cue has already finished playing; retain all of its PCM.\n ctx.postTurnBeginFirstPacketAt = null;\n ctx.postTurnBeginFirstAudioAt = null;\n ctx.turnBegins.push({ atMs: Math.round(at - origin), turn: turn ?? null });\n}\n\nexport function recordUnheardCancellation(ctx, event, at, origin, eventIndex) {\n const prior = ctx.turnBegins.at(-1);\n if (event.event !== \"flush\" || event.reason !== \"turn_merge\"\n || !Number.isSafeInteger(event.cancelled_turn) || event.cancelled_turn <= 0\n || prior?.turn !== event.cancelled_turn || at < ctx.latestTurnBeginAt\n || ctx.samples !== 0 || ctx.assistantTranscriptEvents.length\n || ctx.tools.length || ctx.toolResults.length\n || ctx.unheardCancellations.some(c => c.turnBeginIndex === ctx.turnBegins.length - 1)) return;\n ctx.unheardCancellations.push({ turn: event.cancelled_turn, at,\n atMs: Math.round(at - origin), eventIndex, turnBeginIndex: ctx.turnBegins.length - 1,\n admitted: false });\n}\n\nexport function recordAudio(ctx, pcm, at, rate) {\n if (!(pcm instanceof Int16Array) || !pcm.length) return null;\n const copy = pcm.slice(); // Some transports reuse callback buffers.\n const playbackAt = Math.max(at, ctx.playbackEndAt ?? at);\n const endAt = playbackAt + (copy.length / rate) * 1000;\n const bounds = audibleBounds(copy, rate);\n ctx.firstPacketAt ??= at;\n ctx.lastPacketAt = at;\n ctx.playbackEndAt = endAt;\n const afterTurnBegin = ctx.latestTurnBeginAt != null && at >= ctx.latestTurnBeginAt;\n if (afterTurnBegin) ctx.postTurnBeginFirstPacketAt ??= at;\n if (bounds.first != null) {\n ctx.firstAudioAt ??= playbackAt + (bounds.first / rate) * 1000;\n ctx.lastAudioAt = playbackAt + (bounds.last / rate) * 1000;\n if (afterTurnBegin) ctx.postTurnBeginFirstAudioAt ??= playbackAt + (bounds.first / rate) * 1000;\n }\n ctx.samples += copy.length;\n ctx.pcm.push(copy);\n return { playbackAt, endAt, pcm: copy };\n}\n\n/** Preserve negative gaps: they are overlaps, never missing measurements. */\nexport function captureTiming(ctx, caller, origin = 0) {\n const relative = (at) => at == null ? null : Math.round(at - origin);\n const gap = (at) => at == null || caller.speechOffsetAt == null\n ? null : Math.round(at - caller.speechOffsetAt);\n return {\n ttfbMs: gap(ctx.firstAudioAt),\n anyAudioTtfbMs: gap(ctx.firstAudioAt),\n postTurnBeginTtfbMs: gap(ctx.postTurnBeginFirstAudioAt),\n turnMs: gap(ctx.lastAudioAt),\n packetTtfbMs: gap(ctx.firstPacketAt),\n callerStartMs: relative(caller.startedAt),\n callerSpeechOnsetMs: relative(caller.speechOnsetAt),\n callerSpeechOffsetMs: relative(caller.speechOffsetAt),\n callerStreamEndMs: relative(caller.streamEndAt),\n agentFirstPacketMs: relative(ctx.firstPacketAt),\n agentLastPacketMs: relative(ctx.lastPacketAt),\n agentSpeechOnsetMs: relative(ctx.firstAudioAt),\n agentSpeechOffsetMs: relative(ctx.lastAudioAt),\n agentPlaybackEndMs: relative(ctx.playbackEndAt),\n latestTurnBeginMs: relative(ctx.latestTurnBeginAt),\n postTurnBeginFirstPacketMs: relative(ctx.postTurnBeginFirstPacketAt),\n postTurnBeginSpeechOnsetMs: relative(ctx.postTurnBeginFirstAudioAt),\n lastAssistantTranscriptMs: relative(ctx.lastAssistantTranscriptAt),\n callerMaxFrameGapMs: Math.round(caller.maxFrameGapMs ?? 0),\n method: \"client-monotonic-queued-playback-energy-bounds\",\n };\n}\n\nexport async function streamPcmRealtime(omni, pcm, rate, clock, onFrame = () => {}) {\n const frameSize = Math.max(1, Math.round((rate * FRAME_MS) / 1000));\n const bounds = audibleBounds(pcm, rate);\n const startedAt = clock.now();\n let speechOnsetAt = null;\n let speechOffsetAt = null;\n let previousFrameEnd = startedAt;\n let maxFrameGapMs = 0;\n for (let off = 0; off < pcm.length; off += frameSize) {\n const frame = pcm.subarray(off, Math.min(off + frameSize, pcm.length));\n const at = clock.now();\n maxFrameGapMs = Math.max(maxFrameGapMs, at - previousFrameEnd);\n const durationMs = (frame.length / rate) * 1000;\n omni.sendAudio(frame);\n onFrame(frame, at);\n if (bounds.first != null && bounds.first >= off && bounds.first < off + frame.length) {\n speechOnsetAt = at + ((bounds.first - off) / rate) * 1000;\n }\n if (bounds.last != null && bounds.last > off && bounds.last <= off + frame.length) {\n speechOffsetAt = at + ((bounds.last - off) / rate) * 1000;\n }\n previousFrameEnd = at + durationMs;\n // Pace from the actual send. Never burst late frames to \"catch up\".\n await clock.sleep(Math.max(0, previousFrameEnd - clock.now()));\n }\n return { startedAt, speechOnsetAt, speechOffsetAt,\n streamEndAt: clock.now(), maxFrameGapMs };\n}\n\nexport async function streamSilenceWhile(omni, rate, shouldContinue, clock, onFrame = () => {}) {\n const silence = new Int16Array(Math.max(1, Math.round((rate * FRAME_MS) / 1000)));\n while (shouldContinue()) {\n const at = clock.now();\n omni.sendAudio(silence);\n onFrame(silence, at);\n await clock.sleep(FRAME_MS);\n }\n}\n\nexport async function waitForAgentSettle(getCtx, clock, opts = {}) {\n const started = clock.now();\n const observation = opts.callerStreamEndAt != null;\n if (observation && (!Number.isFinite(opts.callerStreamEndAt)\n || opts.callerStreamEndAt < 0 || opts.callerStreamEndAt > started)) {\n throw new Error(\"Invalid caller observation boundary\");\n }\n const timeoutMs = observation ? INTERRUPTION_OBSERVATION.maximumAfterCallerStreamMs : opts.timeoutMs ?? 25000;\n const settleMs = observation ? INTERRUPTION_OBSERVATION.quietMs : opts.settleMs ?? 2000;\n const deadlineAt = (observation ? opts.callerStreamEndAt : started) + timeoutMs;\n const minimumAt = observation ? opts.callerStreamEndAt + INTERRUPTION_OBSERVATION.minimumAfterCallerStreamMs : -Infinity;\n while (true) {\n const ctx = getCtx();\n const at = clock.now();\n if (opts.isClosed?.()) return { reason: \"closed\", at };\n if (observation && at >= deadlineAt) return { reason: \"timeout\", at };\n const hasResponseAudio = !opts.requireTurnBegin || (ctx.latestTurnBeginAt != null\n && ctx.postTurnBeginFirstAudioAt != null);\n // Synthesis text is progress, not a completion event. It can arrive after\n // early/final audio, so extend the quiet floor without demanding another\n // packet after every advisory. The protocol still has no reply-end marker.\n const activityEnd = Math.max(ctx.playbackEndAt ?? -Infinity,\n ctx.latestTurnBeginAt ?? -Infinity, ctx.lastAssistantTranscriptAt ?? -Infinity,\n observation ? ctx.lastCallerTranscriptAt ?? -Infinity : -Infinity);\n // Do not demand a new post-caller turn_begin: an earlier turn can be\n // coalesced and speak later. This records a bounded quiet observation;\n // downstream response-timing checks can impose a stricter begin boundary.\n if (ctx.firstPacketAt != null && hasResponseAudio && at >= minimumAt && at >= activityEnd + settleMs) {\n return { reason: \"settled\", at };\n }\n if (opts.allowEmpty && ctx.firstPacketAt == null && at >= minimumAt && at - started >= (opts.emptyWaitMs ?? 800)) {\n return { reason: \"empty\", at };\n }\n if (at - started >= timeoutMs) return { reason: \"timeout\", at };\n await clock.sleep(observation ? Math.min(20, deadlineAt - at) : 20);\n }\n}\n\nexport async function withTimeout(promise, ms, message, clock) {\n let timer;\n try {\n return await Promise.race([\n promise,\n new Promise((_, reject) => {\n timer = clock.setTimeout(() => reject(new Error(message)), ms);\n }),\n ]);\n } finally {\n clock.clearTimeout(timer);\n }\n}\n",
|
|
6
6
|
"main.mjs": "import PyAI from \"@pyai/sdk\";\nimport WebSocket from \"ws\"; // Explicit transport makes this work on Node 20 too.\nimport { readFile, writeFile, mkdtemp } from \"node:fs/promises\";\nimport { join } from \"node:path\";\nimport { runSession } from \"./session.mjs\";\nimport { readWav, writeWav, assessAnswer } from \"./audio.mjs\";\n\nasync function main() {\n const [callerPath, interruptionPath, extra] = process.argv.slice(2);\n if (!callerPath || extra) throw new Error(\"Usage: node main.mjs caller.wav [interruption.wav]\");\n if (!process.env.PYAI_API_KEY) throw new Error(\"Inject PYAI_API_KEY through your environment first\");\n const caller = readWav(await readFile(callerPath));\n const interruption = interruptionPath ? readWav(await readFile(interruptionPath)) : undefined;\n const pyai = new PyAI({ apiKey: process.env.PYAI_API_KEY, maxRetries: 0,\n fetch: (url, options) => fetch(url, { ...options, signal: AbortSignal.timeout(30000) }),\n });\n const { report, audio, replyAudio } = await runSession({ pyai, webSocket: WebSocket, caller, interruption });\n const directory = await mkdtemp(join(process.cwd(), \"omni-run-\"));\n const save = (name, data) => writeFile(join(directory, name), data, { mode: 0o600, flag: \"wx\" });\n await save(\"session.wav\", writeWav(audio, report.output_rate ?? 24000));\n const reply = writeWav(replyAudio, report.output_rate ?? 24000);\n await save(\"reply.wav\", reply);\n // Hear the captured bytes, rather than trusting synthesis-advisory text.\n let capturedTranscript = \"\";\n if (report.reply_audio_bytes) {\n try {\n const heard = await pyai.audio.transcriptions.create({ file: new File([reply], \"reply.wav\", { type: \"audio/wav\" }), model: \"pyai-hear\" });\n capturedTranscript = heard.text;\n } catch { report.hear_error = \"Captured audio saved; Hear verification failed\"; }\n }\n const result = { ...report, captured_transcript: capturedTranscript, evidence: assessAnswer(report, capturedTranscript) };\n await save(\"report.json\", JSON.stringify(result, null, 2) + \"\\n\");\n console.log(JSON.stringify({ directory, ...result }, null, 2));\n if (result.evidence.captured_answer !== \"verified_by_hear\" || (interruption && result.evidence.interruption !== \"simulated_queue_cleared\")) process.exitCode = 1;\n}\n\nmain().catch(error => {\n const message = String(error.message).replaceAll(process.env.PYAI_API_KEY || \"\\0\", \"[REDACTED]\");\n console.error(JSON.stringify({ error: message, ...(error.omni_code ? { code: error.omni_code } : {}) })); process.exitCode = 1;\n});\n",
|
|
7
7
|
"session.mjs": "import { newAudioCapture, recordAudio, recordTurnBegin, streamPcmRealtime, INTERRUPTION_OBSERVATION } from \"./timing.mjs\";\n\nexport const realClock = {\n now: () => performance.now(),\n sleep: ms => new Promise(resolve => setTimeout(resolve, ms)),\n};\n\n// Only recognized machine codes may leave the example. Error messages, call\n// IDs and arbitrary code values can contain credentials or customer content.\nconst publicErrorCodes = new Set([\n \"invalid_configure\", \"unsupported_tool_transport\", \"tool_configuration_locked\",\n \"media_dead\", \"unauthorized\", \"insufficient_scope\", \"rate_limit_exceeded\",\n \"concurrency_limit_exceeded\", \"daily_cap_exceeded\", \"credit_exhausted\",\n]);\nconst transportErrorCodes = new Set([\"ECONNRESET\", \"ECONNREFUSED\", \"ETIMEDOUT\", \"ENOTFOUND\", \"EAI_AGAIN\"]);\nexport function omniFailure(error) {\n const code = error instanceof Error\n ? (transportErrorCodes.has(error.code) ? error.code : \"transport_or_protocol_error\")\n : (publicErrorCodes.has(error?.code) ? error.code : \"unrecognized_server_error\");\n return Object.assign(new Error(`Omni failed (${code}); no automatic retry was attempted`), { omni_code: code });\n}\n\n/** One bounded, server-side example. Its playback queue is simulated. */\nexport async function runSession({ pyai, webSocket, caller, interruption, clock = realClock, timeoutMs = 60000 }) {\n const rate = 24000, frameSize = 480, silence = new Int16Array(frameSize);\n if (!(caller instanceof Int16Array) || !caller.length || caller.length > rate * 20) throw new Error(\"Provide 1–20 seconds of caller PCM16 at 24 kHz\");\n if (interruption && (!(interruption instanceof Int16Array) || !interruption.length || interruption.length > rate * 20)) throw new Error(\"Invalid interruption PCM\");\n const started = clock.now(), all = newAudioCapture(started), reply = newAudioCapture(started);\n const report = { input_rate: rate, output_rate: null, playback_sink: \"simulated\", tool_calls: 0, tool_executions: 0, duplicate_tool_calls: 0, flushes_after_interruption: 0, cleared_queue_ms: 0, interruption_requested: !!interruption, caller_started_ms: null, interruption_started_ms: null, advisory_transcripts: [], end_reason: \"timeout\" };\n let session, configuredAt = null, closed = false, failure = null, phase = \"greeting\", offset = 0, queueEnd = started, interruptionSent = false, inputEndAt = started;\n const handled = new Set();\n const fail = message => { failure ??= new Error(message); };\n try {\n session = pyai.omni.connect({\n webSocket, rate,\n configure: {\n voice_id: \"stock_amos_en_us\", language: \"en\", greeting: \"Hello. How can I help?\",\n persona: \"When asked about office hours, call lookup_office_hours. Never guess. After the result, say one short sentence giving its opening_time.\",\n tools: [{ name: \"lookup_office_hours\", description: \"Read the synthetic example office opening time.\", side_effect: \"read\", parameters: { type: \"object\", properties: {}, required: [], additionalProperties: false } }],\n },\n onHello: frame => {\n const match = /^pcm16@(8000|24000)$/.exec(frame.audio_out ?? \"\");\n if (!match) return fail(\"Unsupported Omni output format\");\n report.output_rate = Number(match[1]);\n },\n onConfigured: () => { configuredAt ??= clock.now(); },\n onAudio: chunk => {\n if (!report.output_rate) return fail(\"Audio arrived before its output format\");\n const bytes = chunk instanceof ArrayBuffer ? new Uint8Array(chunk) : new Uint8Array(chunk.buffer, chunk.byteOffset, chunk.byteLength);\n if (bytes.byteLength % 2) return fail(\"Invalid PCM16 output\");\n const view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);\n const pcm = Int16Array.from({ length: bytes.length / 2 }, (_, i) => view.getInt16(i * 2, true));\n const now = clock.now();\n recordAudio(all, pcm, now, report.output_rate);\n queueEnd = Math.max(now, queueEnd) + pcm.length / report.output_rate * 1000;\n if (report.tool_executions) recordAudio(reply, pcm, now, report.output_rate);\n },\n onEvent: frame => {\n if (frame.event === \"turn_begin\") recordTurnBegin(all, clock.now(), frame.turn);\n },\n onTranscript: frame => {\n // An assistant advisory is progress; never treat it as received speech.\n if (report.advisory_transcripts.length < 100) report.advisory_transcripts.push({ role: frame.role, text: frame.text });\n if (frame.role === \"assistant\") all.lastAssistantTranscriptAt = clock.now();\n },\n onToolCall: frame => {\n report.tool_calls++;\n if (handled.has(frame.call_id)) { report.duplicate_tool_calls++; return; }\n handled.add(frame.call_id);\n if (frame.name !== \"lookup_office_hours\") { session.toolResult(frame.call_id, { error: \"Unknown example tool\" }); return; }\n // Replace this read-only fixture with your authorized data lookup.\n report.tool_executions++;\n session.toolResult(frame.call_id, { result: { opening_time: \"9 a.m.\", source: \"synthetic example office\" } });\n },\n onBargeIn: () => {\n const now = clock.now(), queued = Math.max(0, queueEnd - now);\n if (report.interruption_started_ms != null) { report.flushes_after_interruption++; report.cleared_queue_ms += queued; }\n // A real speaker adapter must cancel scheduled and currently playing audio.\n queueEnd = now; all.playbackEndAt = now; reply.playbackEndAt = now;\n },\n onError: error => { failure ??= omniFailure(error); },\n onClose: () => { closed = true; },\n });\n while (clock.now() - started < timeoutMs) {\n if (failure) throw failure;\n if (closed) { report.end_reason = \"closed\"; break; }\n const now = clock.now();\n if (configuredAt == null || !report.output_rate) {\n if (now - started >= 10000) throw new Error(\"Omni connection/configuration timed out\");\n await clock.sleep(20); continue;\n }\n const quietAt = Math.max(queueEnd, all.lastPacketAt ?? now, all.lastAssistantTranscriptAt ?? 0, inputEndAt);\n if (phase === \"greeting\") {\n if (all.firstPacketAt != null && now - configuredAt >= 2000 && now >= quietAt + 800) {\n phase = \"caller\"; report.caller_started_ms = now - started;\n } else if (now - configuredAt > 15000) throw new Error(\"Greeting did not drain within 15 seconds\");\n }\n if (phase === \"reply\" && interruption && !interruptionSent && reply.samples && queueEnd - now >= 200) {\n phase = \"interruption\"; offset = 0; interruptionSent = true; report.interruption_started_ms = now - started;\n }\n const observationEnd = inputEndAt + (interruptionSent ? INTERRUPTION_OBSERVATION.minimumAfterCallerStreamMs : 0);\n if (phase === \"reply\" && reply.samples && now >= quietAt + 2000 && now >= observationEnd && (!interruption || interruptionSent)) {\n report.end_reason = \"quiet_window\"; break;\n }\n // Exactly one producer: each 20 ms slot contains caller PCM OR silence.\n let pcm = silence;\n if (phase === \"caller\" || phase === \"interruption\") {\n const source = phase === \"caller\" ? caller : interruption;\n pcm = source.subarray(offset, offset + frameSize); offset += pcm.length;\n if (offset >= source.length) {\n inputEndAt = now + pcm.length / rate * 1000;\n report[phase === \"caller\" ? \"caller_completed_ms\" : \"interruption_completed_ms\"] = inputEndAt - started;\n phase = \"reply\"; offset = 0;\n }\n }\n await streamPcmRealtime(session, pcm, rate, clock);\n }\n report.audio_bytes = all.samples * 2;\n report.reply_audio_bytes = reply.samples * 2;\n report.elapsed_ms = Math.round(clock.now() - started);\n return { report, audio: all.pcm, replyAudio: reply.pcm };\n } finally { session?.close(); }\n}\n",
|
package/dist/dialer/index.d.ts
CHANGED
|
@@ -2,6 +2,7 @@ export * from "./managedDialer.ts";
|
|
|
2
2
|
export * from "./dialingQueue.ts";
|
|
3
3
|
export * from "./indexedDbJournal.ts";
|
|
4
4
|
export * from "./voiceSdkAudio.ts";
|
|
5
|
+
export * from "./managedBrowserAudio.ts";
|
|
5
6
|
export * from "./meetingRecorder.ts";
|
|
6
7
|
export * from "./meetingRecordingStore.ts";
|
|
7
8
|
export * from "./meetingIntelligence.ts";
|
package/dist/dialer/index.js
CHANGED
|
@@ -2,6 +2,7 @@ export * from "./managedDialer.js";
|
|
|
2
2
|
export * from "./dialingQueue.js";
|
|
3
3
|
export * from "./indexedDbJournal.js";
|
|
4
4
|
export * from "./voiceSdkAudio.js";
|
|
5
|
+
export * from "./managedBrowserAudio.js";
|
|
5
6
|
export * from "./meetingRecorder.js";
|
|
6
7
|
export * from "./meetingRecordingStore.js";
|
|
7
8
|
export * from "./meetingIntelligence.js";
|
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
import type { DialerAudio } from "./managedDialer.ts";
|
|
2
|
+
import type { TelephonyBrowserSession } from "../index.ts";
|
|
3
|
+
/** Issued by the application's server after PyAI has authorized one call.
|
|
4
|
+
* This contains no workspace key or reusable carrier credential. */
|
|
5
|
+
export type BrowserAudioCapability = Pick<TelephonyBrowserSession, "access_token" | "connect_params"> & {
|
|
6
|
+
transport?: "voice-sdk";
|
|
7
|
+
};
|
|
8
|
+
type AudioEvent = Parameters<DialerAudio["connect"]>[1];
|
|
9
|
+
export interface ManagedBrowserAudio {
|
|
10
|
+
connect(capability: BrowserAudioCapability, event: AudioEvent): Promise<void>;
|
|
11
|
+
disconnect(): void;
|
|
12
|
+
mute(value: boolean): void;
|
|
13
|
+
sendDigits(digits: string): void;
|
|
14
|
+
}
|
|
15
|
+
export interface ManagedBrowserAudioOptions {
|
|
16
|
+
/** Optional media adapter for embedded hosts. Ordinary applications omit it. */
|
|
17
|
+
loadTransport?: (transport: "voice-sdk") => Promise<DialerAudio>;
|
|
18
|
+
}
|
|
19
|
+
/** Provider-owned, lazy media setup. A cancelled import cannot place a late call. */
|
|
20
|
+
export declare function createManagedBrowserAudio(options?: ManagedBrowserAudioOptions): ManagedBrowserAudio;
|
|
21
|
+
export {};
|
|
@@ -0,0 +1,69 @@
|
|
|
1
|
+
import { VoiceSdkAudio } from "./voiceSdkAudio.js";
|
|
2
|
+
async function loadTransport() {
|
|
3
|
+
// The upstream transport is an implementation detail of this PyAI adapter.
|
|
4
|
+
// Applications must not import or construct the carrier's Device themselves.
|
|
5
|
+
const { Device } = await import("@twilio/voice-sdk");
|
|
6
|
+
return new VoiceSdkAudio(token => new Device(token));
|
|
7
|
+
}
|
|
8
|
+
/** Provider-owned, lazy media setup. A cancelled import cannot place a late call. */
|
|
9
|
+
export function createManagedBrowserAudio(options = {}) {
|
|
10
|
+
let generation = 0;
|
|
11
|
+
let audio = null;
|
|
12
|
+
const disconnect = () => {
|
|
13
|
+
generation++;
|
|
14
|
+
const prior = audio;
|
|
15
|
+
audio = null;
|
|
16
|
+
try {
|
|
17
|
+
prior?.disconnect();
|
|
18
|
+
}
|
|
19
|
+
catch { /* Remote call cleanup belongs to the server broker. */ }
|
|
20
|
+
};
|
|
21
|
+
return {
|
|
22
|
+
async connect(capability, event) {
|
|
23
|
+
const params = capability.connect_params;
|
|
24
|
+
if (!capability.access_token || capability.transport !== undefined && capability.transport !== "voice-sdk" ||
|
|
25
|
+
!params || Object.keys(params).length !== 1 || typeof params.session_id !== "string" ||
|
|
26
|
+
params.session_id.length < 1 || params.session_id.length > 200) {
|
|
27
|
+
throw new Error("A supported, single-call PyAI audio capability is required.");
|
|
28
|
+
}
|
|
29
|
+
// Capture before awaiting a lazy import; never retain a mutable caller object.
|
|
30
|
+
const session = { access_token: capability.access_token, connect_params: { session_id: params.session_id } };
|
|
31
|
+
disconnect();
|
|
32
|
+
const current = generation;
|
|
33
|
+
const candidate = await (options.loadTransport ?? loadTransport)("voice-sdk");
|
|
34
|
+
if (current !== generation) {
|
|
35
|
+
try {
|
|
36
|
+
candidate.disconnect();
|
|
37
|
+
}
|
|
38
|
+
catch { /* Not connected. */ }
|
|
39
|
+
return;
|
|
40
|
+
}
|
|
41
|
+
audio = candidate;
|
|
42
|
+
try {
|
|
43
|
+
// DialerAudio implementations consume this same opaque media capability.
|
|
44
|
+
await candidate.connect(session, state => {
|
|
45
|
+
if (current === generation)
|
|
46
|
+
event(state);
|
|
47
|
+
});
|
|
48
|
+
}
|
|
49
|
+
catch (error) {
|
|
50
|
+
if (current === generation)
|
|
51
|
+
disconnect();
|
|
52
|
+
throw error;
|
|
53
|
+
}
|
|
54
|
+
},
|
|
55
|
+
disconnect,
|
|
56
|
+
mute(value) {
|
|
57
|
+
if (!audio)
|
|
58
|
+
throw new Error("No active audio.");
|
|
59
|
+
audio.mute(value);
|
|
60
|
+
},
|
|
61
|
+
sendDigits(digits) {
|
|
62
|
+
if (!audio)
|
|
63
|
+
throw new Error("No active audio.");
|
|
64
|
+
if (!/^[0-9*#]{1,32}$/.test(digits))
|
|
65
|
+
throw new Error("Use one to 32 keypad digits.");
|
|
66
|
+
audio.sendDigits(digits);
|
|
67
|
+
},
|
|
68
|
+
};
|
|
69
|
+
}
|
|
@@ -5,6 +5,7 @@ export interface DialerBroker {
|
|
|
5
5
|
create(input: TelephonyBrowserSessionCreate, idempotencyKey: string): Promise<TelephonyBrowserSession>;
|
|
6
6
|
get(callId: string): Promise<TelephonyCall>;
|
|
7
7
|
hangup(callId: string): Promise<TelephonyCall>;
|
|
8
|
+
setHold?(callId: string, held: boolean, key: string): Promise<TelephonyCall>;
|
|
8
9
|
transfers?: {
|
|
9
10
|
create(callId: string, input: {
|
|
10
11
|
to: string;
|
|
@@ -25,6 +25,13 @@ export function createServerDialerBroker(sdk, authenticatedUserId) {
|
|
|
25
25
|
return call;
|
|
26
26
|
return sdk.telephony.calls.hangup(id);
|
|
27
27
|
},
|
|
28
|
+
async setHold(id, held, key) {
|
|
29
|
+
await ownedCall(id);
|
|
30
|
+
if (!key || key.length > 255 || /[\r\n]/.test(key))
|
|
31
|
+
throw new Error("A valid hold retry key is required.");
|
|
32
|
+
const scopedKey = "dialer-hold:" + createHash("sha256").update(authenticatedUserId).update("\0").update(id).update("\0").update(key).digest("hex");
|
|
33
|
+
return sdk.telephony.calls.setHold(id, { held, user_id: authenticatedUserId }, { idempotencyKey: scopedKey });
|
|
34
|
+
},
|
|
28
35
|
transfers: {
|
|
29
36
|
async create(id, input, key) {
|
|
30
37
|
await ownedCall(id);
|
package/dist/index.d.ts
CHANGED
|
@@ -857,6 +857,8 @@ export interface TelephonyCall {
|
|
|
857
857
|
agent_id: string | null;
|
|
858
858
|
call_type?: "human";
|
|
859
859
|
transfer_enabled?: boolean;
|
|
860
|
+
held?: boolean;
|
|
861
|
+
hold_pending?: boolean;
|
|
860
862
|
transfers?: TelephonyTransfer[];
|
|
861
863
|
user_id?: string;
|
|
862
864
|
client_reference: string | null;
|
|
@@ -874,6 +876,37 @@ export interface TelephonyTransfer {
|
|
|
874
876
|
status: "holding" | "dialing" | "ready" | "completing" | "canceling" | "completed" | "canceled" | "failed";
|
|
875
877
|
to: string | null;
|
|
876
878
|
}
|
|
879
|
+
export interface TelephonyCallQuality {
|
|
880
|
+
call_id: string;
|
|
881
|
+
status: "available" | "unavailable";
|
|
882
|
+
reason: string | null;
|
|
883
|
+
source: "managed_transport" | null;
|
|
884
|
+
scope: "caller_to_platform" | null;
|
|
885
|
+
observation: "terminal_snapshot" | "partial_snapshot" | null;
|
|
886
|
+
observed_at: string | null;
|
|
887
|
+
live: false;
|
|
888
|
+
metrics: Record<"packets_received" | "packets_lost" | "packets_reordered" | "duplicate_packets" | "packets_sent" | "playout_frames" | "concealed_frames" | "late_frames" | "playout_duplicates" | "jitter_ms" | "packet_loss_percent" | "concealment_percent", number | null> & {
|
|
889
|
+
mos: null;
|
|
890
|
+
rtt_ms: null;
|
|
891
|
+
};
|
|
892
|
+
unavailable_metrics: string[];
|
|
893
|
+
}
|
|
894
|
+
export interface TelephonyCallCapabilities {
|
|
895
|
+
object: "telephony.call_capabilities";
|
|
896
|
+
call_id: string;
|
|
897
|
+
version: 1;
|
|
898
|
+
call_type: "human" | "ai";
|
|
899
|
+
controls: Record<"hangup" | "hold" | "warm_transfer" | "cold_transfer" | "complete_transfer" | "cancel_transfer" | "listen" | "whisper" | "barge", {
|
|
900
|
+
supported: boolean;
|
|
901
|
+
reason: string | null;
|
|
902
|
+
}>;
|
|
903
|
+
client_controls: Array<"mute" | "dtmf">;
|
|
904
|
+
quality: {
|
|
905
|
+
read_supported: true;
|
|
906
|
+
live: false;
|
|
907
|
+
};
|
|
908
|
+
revalidation_required: true;
|
|
909
|
+
}
|
|
877
910
|
export interface TelephonyCallCreate {
|
|
878
911
|
from_number_id: string;
|
|
879
912
|
to: string;
|
|
@@ -1507,6 +1540,8 @@ export declare class PyAI {
|
|
|
1507
1540
|
idempotencyKey: string;
|
|
1508
1541
|
}) => Promise<TelephonyCall>;
|
|
1509
1542
|
get: (id: string) => Promise<TelephonyCall>;
|
|
1543
|
+
capabilities: (id: string) => Promise<TelephonyCallCapabilities>;
|
|
1544
|
+
quality: (id: string) => Promise<TelephonyCallQuality>;
|
|
1510
1545
|
list: (input?: {
|
|
1511
1546
|
limit?: number;
|
|
1512
1547
|
cursor?: string;
|
|
@@ -1526,6 +1561,13 @@ export declare class PyAI {
|
|
|
1526
1561
|
range?: string;
|
|
1527
1562
|
}) => Promise<Response>;
|
|
1528
1563
|
deleteRecording: (id: string) => Promise<void>;
|
|
1564
|
+
/** The same durable key recovers an uncertain hold command without repeating it. */
|
|
1565
|
+
setHold: (id: string, input: {
|
|
1566
|
+
user_id: string;
|
|
1567
|
+
held: boolean;
|
|
1568
|
+
}, options: {
|
|
1569
|
+
idempotencyKey: string;
|
|
1570
|
+
}) => Promise<TelephonyCall>;
|
|
1529
1571
|
transfers: {
|
|
1530
1572
|
/** Persist the key and target before requesting a consultation. */
|
|
1531
1573
|
create: (id: string, input: {
|
package/dist/index.js
CHANGED
|
@@ -1183,6 +1183,8 @@ export class PyAI {
|
|
|
1183
1183
|
calls: {
|
|
1184
1184
|
create: (input, options) => this.postJson("/v1/telephony/calls", input, telephonyIdempotencyHeader(options.idempotencyKey)),
|
|
1185
1185
|
get: (id) => this.getJson(`/v1/telephony/calls/${encodeURIComponent(id)}`),
|
|
1186
|
+
capabilities: (id) => this.getJson(`/v1/telephony/calls/${encodeURIComponent(id)}/capabilities`),
|
|
1187
|
+
quality: (id) => this.getJson(`/v1/telephony/calls/${encodeURIComponent(id)}/quality`),
|
|
1186
1188
|
list: (input = {}) => {
|
|
1187
1189
|
const query = new URLSearchParams();
|
|
1188
1190
|
for (const [key, value] of Object.entries(input))
|
|
@@ -1195,6 +1197,8 @@ export class PyAI {
|
|
|
1195
1197
|
decideAdmission: (id, input) => this.postJson(`/v1/telephony/calls/${encodeURIComponent(id)}/admission`, input),
|
|
1196
1198
|
recording: (id, options = {}) => this.request(`/v1/telephony/calls/${encodeURIComponent(id)}/recording`, { headers: this.authHeaders(options.range ? { Range: options.range } : {}) }),
|
|
1197
1199
|
deleteRecording: async (id) => { await this.deleteReq(`/v1/telephony/calls/${encodeURIComponent(id)}/recording`); },
|
|
1200
|
+
/** The same durable key recovers an uncertain hold command without repeating it. */
|
|
1201
|
+
setHold: (id, input, options) => this.postJson(`/v1/telephony/calls/${encodeURIComponent(id)}/hold`, input, telephonyIdempotencyHeader(options.idempotencyKey)),
|
|
1198
1202
|
transfers: {
|
|
1199
1203
|
/** Persist the key and target before requesting a consultation. */
|
|
1200
1204
|
create: (id, input, options) => this.postJson(`/v1/telephony/calls/${encodeURIComponent(id)}/transfers`, input, telephonyIdempotencyHeader(options.idempotencyKey)),
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@pyai/sdk",
|
|
3
|
-
"version": "0.
|
|
3
|
+
"version": "0.16.0",
|
|
4
4
|
"description": "Official TypeScript/JavaScript SDK for PyAI, speech-to-text (Hear), text-to-speech (Speak), realtime voice agents (Omni), and call compliance (Trace).",
|
|
5
5
|
"type": "module",
|
|
6
6
|
"main": "dist/index.js",
|
package/src/cli-omni-template.ts
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
// Generated by scripts/generate-omni-starter.mjs; edit examples/omni-starter instead.
|
|
2
2
|
export const omniTemplate = Object.freeze({
|
|
3
|
-
"README.md": "# Runnable Omni starter\n\nRequires Node 20.19+ and ESM. Node 20 uses the included `ws` transport;\nNode 22 also works. The separate MCP server requires Node 22+.\nThis server-side example uses a synthetic read-only office-hours tool. Read\n`PYAI.md` when generating the project through the CLI.\n\n<!-- omni-install:start -->\n```sh\nnpm install @pyai/sdk@0.
|
|
4
|
-
"package.json": "{\n \"name\": \"pyai-omni-starter\",\n \"version\": \"0.0.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20.19.0\"\n },\n \"scripts\": {\n \"start\": \"node main.mjs\"\n },\n \"dependencies\": {\n \"@pyai/sdk\": \"0.
|
|
3
|
+
"README.md": "# Runnable Omni starter\n\nRequires Node 20.19+ and ESM. Node 20 uses the included `ws` transport;\nNode 22 also works. The separate MCP server requires Node 22+.\nThis server-side example uses a synthetic read-only office-hours tool. Read\n`PYAI.md` when generating the project through the CLI.\n\n<!-- omni-install:start -->\n```sh\nnpm install @pyai/sdk@0.16.0\nnpx pyai init voice-demo --template omni\ncd voice-demo\nnpm install\n```\n\nInitialization is offline; installing dependencies and running `main.mjs` are\nseparate steps. For an existing checkout of this example, run `npm install`\nin its directory instead.\n<!-- omni-install:end -->\n\nInject `PYAI_API_KEY` through your secret manager or shell environment. Do not\npaste a real key into source, a prompt, or shell history. `.env.example` documents\nthe variable; this program does not load `.env` or saved CLI profiles.\n\nPrepare a 24 kHz PCM16 mono WAV, at most 20 seconds, saying “Please look up the\noffice opening time.” Then run:\n\n```sh\nnode main.mjs caller.wav\n```\n\nRunning the program explicitly opens one bounded Omni session and uploads its\ncaptured reply to Hear. Both consume your key's allowed usage. Automatic retries\nare disabled. No managed number is bought and no phone call is placed.\n\n`session.mjs` waits for configuration and the greeting to drain, sends exactly\none paced input stream (caller PCM or silence), returns the read tool's result,\nand captures output using the rate from `hello.audio_out`. Never run a separate\nsilence timer alongside caller audio. `timing.mjs` is generated from the existing\nOmni evaluation harness's timing utilities.\n\nThe new private `omni-run-*` folder contains `session.wav`, `reply.wav`, and\n`report.json`. The report separates received audio, an office-hours answer\nrecovered by Hear from captured audio, and physical playback (not tested).\nThe office-hours check accepts formatting such as `opens@9a.m`, `9 a.m.` and\n`nine in the morning` in a complete affirmative opening-time sentence. Wrong\ntimes, negation, conflicting or truncated text remain unverified. This is a\nconservative check for the example fact, not a general answer-quality score.\nA synthesis transcript is advisory; it is not evidence that speech arrived.\nThis starter does not negotiate playback receipts. Capture ends after the\nsimulated queue drains and two seconds of quiet, a bounded capture heuristic.\nListen to the saved file when qualifying actual sound quality or completeness.\n\nTo exercise interruption, supply a second WAV saying “Stop speaking now”:\n\n```sh\nnode main.mjs caller.wav interruption.wav\n```\n\nThe second clip starts while reply audio is queued. `onBargeIn` clears that\nsimulated queue, and capture observes at least five seconds after that clip ends.\nAn interrupted answer can correctly fail the full-answer check;\ninspect the separate interruption evidence. This is not a physical speaker test.\nFor a real playback adapter, cancel both queued and currently playing audio.\nReplace the example lookup with authorized application data before product use.\n\nOn a session error, the program exits with JSON containing a recognized machine\n`code` (for example `media_dead` or `invalid_configure`). Unknown server codes\nbecome `unrecognized_server_error`; transport or decoding errors without a\nrecognized transport code become `transport_or_protocol_error`. Raw error text,\ncall IDs and unknown code values are omitted. No automatic retry is attempted;\ninspect the code and your authorized diagnostics before deciding to retry.\n",
|
|
4
|
+
"package.json": "{\n \"name\": \"pyai-omni-starter\",\n \"version\": \"0.0.0\",\n \"private\": true,\n \"type\": \"module\",\n \"engines\": {\n \"node\": \">=20.19.0\"\n },\n \"scripts\": {\n \"start\": \"node main.mjs\"\n },\n \"dependencies\": {\n \"@pyai/sdk\": \"0.16.0\",\n \"ws\": \"^8.18.3\"\n }\n}\n",
|
|
5
5
|
"timing.mjs": "// Monotonic capture timing. Arrival time is distinct from queued playback:\n// a single packet may contain seconds of audio and must finish before settling.\nexport const FRAME_MS = 20;\nexport const REALTIME_GAP_LIMIT_MS = 100;\n// Interruption tests need an observation window after caller transmission,\n// even when an earlier cue/response already finished. These are test bounds,\n// not an inferred engine reply-end or a naturalness threshold.\nexport const INTERRUPTION_OBSERVATION = Object.freeze({\n minimumAfterCallerStreamMs: 5000, quietMs: 2000, maximumAfterCallerStreamMs: 25000,\n});\n\n/** Approximate audible bounds; this is energy detection, not a speech/VAD claim. */\nexport function audibleBounds(pcm, rate, rmsFloor = 160) {\n const window = Math.max(1, Math.round(rate / 100)); // 10 ms windows\n let first = null;\n let last = null;\n for (let off = 0; off < pcm.length; off += window) {\n const end = Math.min(pcm.length, off + window);\n let energy = 0;\n for (let i = off; i < end; i++) energy += pcm[i] * pcm[i];\n if (Math.sqrt(energy / (end - off)) >= rmsFloor) {\n first ??= off;\n last = end;\n }\n }\n return { first, last };\n}\n\nexport function newAudioCapture(started) {\n return {\n started, firstPacketAt: null, lastPacketAt: null,\n firstAudioAt: null, lastAudioAt: null, playbackEndAt: null,\n samples: 0, pcm: [], turnBeginAt: null, eouMs: null,\n latestTurnBeginAt: null, postTurnBeginFirstPacketAt: null,\n postTurnBeginFirstAudioAt: null, lastAssistantTranscriptAt: null, lastCallerTranscriptAt: null,\n callerTranscriptEvents: [], assistantTranscriptEvents: [],\n turnBegins: [], unheardCancellations: [], unheardSupersededTurns: [], tools: [], toolResults: [], kb: null, events: [],\n };\n}\n\nexport function recordTurnBegin(ctx, at, turn, origin = 0) {\n const prior = ctx.turnBegins.at(-1);\n const cancellation = ctx.unheardCancellations.at(-1);\n if (prior && cancellation && !cancellation.admitted\n && cancellation.turn === prior.turn && cancellation.turnBeginIndex === ctx.turnBegins.length - 1\n && Number.isSafeInteger(turn) && turn > prior.turn && at > cancellation.at\n && ctx.samples === 0 && ctx.assistantTranscriptEvents.length === 0\n && ctx.tools.length === 0 && ctx.toolResults.length === 0) {\n cancellation.admitted = true;\n ctx.unheardSupersededTurns.push({ turn: prior.turn, beginAtMs: prior.atMs,\n cancelledAtMs: cancellation.atMs, cancellationEventIndex: cancellation.eventIndex,\n replacementTurn: turn, outputPacketsBeforeReplacement: 0 });\n }\n ctx.turnBeginAt ??= at;\n ctx.latestTurnBeginAt = at;\n // Earlier PCM can be a listening cue. A new reply is still pending even\n // when that cue has already finished playing; retain all of its PCM.\n ctx.postTurnBeginFirstPacketAt = null;\n ctx.postTurnBeginFirstAudioAt = null;\n ctx.turnBegins.push({ atMs: Math.round(at - origin), turn: turn ?? null });\n}\n\nexport function recordUnheardCancellation(ctx, event, at, origin, eventIndex) {\n const prior = ctx.turnBegins.at(-1);\n if (event.event !== \"flush\" || event.reason !== \"turn_merge\"\n || !Number.isSafeInteger(event.cancelled_turn) || event.cancelled_turn <= 0\n || prior?.turn !== event.cancelled_turn || at < ctx.latestTurnBeginAt\n || ctx.samples !== 0 || ctx.assistantTranscriptEvents.length\n || ctx.tools.length || ctx.toolResults.length\n || ctx.unheardCancellations.some(c => c.turnBeginIndex === ctx.turnBegins.length - 1)) return;\n ctx.unheardCancellations.push({ turn: event.cancelled_turn, at,\n atMs: Math.round(at - origin), eventIndex, turnBeginIndex: ctx.turnBegins.length - 1,\n admitted: false });\n}\n\nexport function recordAudio(ctx, pcm, at, rate) {\n if (!(pcm instanceof Int16Array) || !pcm.length) return null;\n const copy = pcm.slice(); // Some transports reuse callback buffers.\n const playbackAt = Math.max(at, ctx.playbackEndAt ?? at);\n const endAt = playbackAt + (copy.length / rate) * 1000;\n const bounds = audibleBounds(copy, rate);\n ctx.firstPacketAt ??= at;\n ctx.lastPacketAt = at;\n ctx.playbackEndAt = endAt;\n const afterTurnBegin = ctx.latestTurnBeginAt != null && at >= ctx.latestTurnBeginAt;\n if (afterTurnBegin) ctx.postTurnBeginFirstPacketAt ??= at;\n if (bounds.first != null) {\n ctx.firstAudioAt ??= playbackAt + (bounds.first / rate) * 1000;\n ctx.lastAudioAt = playbackAt + (bounds.last / rate) * 1000;\n if (afterTurnBegin) ctx.postTurnBeginFirstAudioAt ??= playbackAt + (bounds.first / rate) * 1000;\n }\n ctx.samples += copy.length;\n ctx.pcm.push(copy);\n return { playbackAt, endAt, pcm: copy };\n}\n\n/** Preserve negative gaps: they are overlaps, never missing measurements. */\nexport function captureTiming(ctx, caller, origin = 0) {\n const relative = (at) => at == null ? null : Math.round(at - origin);\n const gap = (at) => at == null || caller.speechOffsetAt == null\n ? null : Math.round(at - caller.speechOffsetAt);\n return {\n ttfbMs: gap(ctx.firstAudioAt),\n anyAudioTtfbMs: gap(ctx.firstAudioAt),\n postTurnBeginTtfbMs: gap(ctx.postTurnBeginFirstAudioAt),\n turnMs: gap(ctx.lastAudioAt),\n packetTtfbMs: gap(ctx.firstPacketAt),\n callerStartMs: relative(caller.startedAt),\n callerSpeechOnsetMs: relative(caller.speechOnsetAt),\n callerSpeechOffsetMs: relative(caller.speechOffsetAt),\n callerStreamEndMs: relative(caller.streamEndAt),\n agentFirstPacketMs: relative(ctx.firstPacketAt),\n agentLastPacketMs: relative(ctx.lastPacketAt),\n agentSpeechOnsetMs: relative(ctx.firstAudioAt),\n agentSpeechOffsetMs: relative(ctx.lastAudioAt),\n agentPlaybackEndMs: relative(ctx.playbackEndAt),\n latestTurnBeginMs: relative(ctx.latestTurnBeginAt),\n postTurnBeginFirstPacketMs: relative(ctx.postTurnBeginFirstPacketAt),\n postTurnBeginSpeechOnsetMs: relative(ctx.postTurnBeginFirstAudioAt),\n lastAssistantTranscriptMs: relative(ctx.lastAssistantTranscriptAt),\n callerMaxFrameGapMs: Math.round(caller.maxFrameGapMs ?? 0),\n method: \"client-monotonic-queued-playback-energy-bounds\",\n };\n}\n\nexport async function streamPcmRealtime(omni, pcm, rate, clock, onFrame = () => {}) {\n const frameSize = Math.max(1, Math.round((rate * FRAME_MS) / 1000));\n const bounds = audibleBounds(pcm, rate);\n const startedAt = clock.now();\n let speechOnsetAt = null;\n let speechOffsetAt = null;\n let previousFrameEnd = startedAt;\n let maxFrameGapMs = 0;\n for (let off = 0; off < pcm.length; off += frameSize) {\n const frame = pcm.subarray(off, Math.min(off + frameSize, pcm.length));\n const at = clock.now();\n maxFrameGapMs = Math.max(maxFrameGapMs, at - previousFrameEnd);\n const durationMs = (frame.length / rate) * 1000;\n omni.sendAudio(frame);\n onFrame(frame, at);\n if (bounds.first != null && bounds.first >= off && bounds.first < off + frame.length) {\n speechOnsetAt = at + ((bounds.first - off) / rate) * 1000;\n }\n if (bounds.last != null && bounds.last > off && bounds.last <= off + frame.length) {\n speechOffsetAt = at + ((bounds.last - off) / rate) * 1000;\n }\n previousFrameEnd = at + durationMs;\n // Pace from the actual send. Never burst late frames to \"catch up\".\n await clock.sleep(Math.max(0, previousFrameEnd - clock.now()));\n }\n return { startedAt, speechOnsetAt, speechOffsetAt,\n streamEndAt: clock.now(), maxFrameGapMs };\n}\n\nexport async function streamSilenceWhile(omni, rate, shouldContinue, clock, onFrame = () => {}) {\n const silence = new Int16Array(Math.max(1, Math.round((rate * FRAME_MS) / 1000)));\n while (shouldContinue()) {\n const at = clock.now();\n omni.sendAudio(silence);\n onFrame(silence, at);\n await clock.sleep(FRAME_MS);\n }\n}\n\nexport async function waitForAgentSettle(getCtx, clock, opts = {}) {\n const started = clock.now();\n const observation = opts.callerStreamEndAt != null;\n if (observation && (!Number.isFinite(opts.callerStreamEndAt)\n || opts.callerStreamEndAt < 0 || opts.callerStreamEndAt > started)) {\n throw new Error(\"Invalid caller observation boundary\");\n }\n const timeoutMs = observation ? INTERRUPTION_OBSERVATION.maximumAfterCallerStreamMs : opts.timeoutMs ?? 25000;\n const settleMs = observation ? INTERRUPTION_OBSERVATION.quietMs : opts.settleMs ?? 2000;\n const deadlineAt = (observation ? opts.callerStreamEndAt : started) + timeoutMs;\n const minimumAt = observation ? opts.callerStreamEndAt + INTERRUPTION_OBSERVATION.minimumAfterCallerStreamMs : -Infinity;\n while (true) {\n const ctx = getCtx();\n const at = clock.now();\n if (opts.isClosed?.()) return { reason: \"closed\", at };\n if (observation && at >= deadlineAt) return { reason: \"timeout\", at };\n const hasResponseAudio = !opts.requireTurnBegin || (ctx.latestTurnBeginAt != null\n && ctx.postTurnBeginFirstAudioAt != null);\n // Synthesis text is progress, not a completion event. It can arrive after\n // early/final audio, so extend the quiet floor without demanding another\n // packet after every advisory. The protocol still has no reply-end marker.\n const activityEnd = Math.max(ctx.playbackEndAt ?? -Infinity,\n ctx.latestTurnBeginAt ?? -Infinity, ctx.lastAssistantTranscriptAt ?? -Infinity,\n observation ? ctx.lastCallerTranscriptAt ?? -Infinity : -Infinity);\n // Do not demand a new post-caller turn_begin: an earlier turn can be\n // coalesced and speak later. This records a bounded quiet observation;\n // downstream response-timing checks can impose a stricter begin boundary.\n if (ctx.firstPacketAt != null && hasResponseAudio && at >= minimumAt && at >= activityEnd + settleMs) {\n return { reason: \"settled\", at };\n }\n if (opts.allowEmpty && ctx.firstPacketAt == null && at >= minimumAt && at - started >= (opts.emptyWaitMs ?? 800)) {\n return { reason: \"empty\", at };\n }\n if (at - started >= timeoutMs) return { reason: \"timeout\", at };\n await clock.sleep(observation ? Math.min(20, deadlineAt - at) : 20);\n }\n}\n\nexport async function withTimeout(promise, ms, message, clock) {\n let timer;\n try {\n return await Promise.race([\n promise,\n new Promise((_, reject) => {\n timer = clock.setTimeout(() => reject(new Error(message)), ms);\n }),\n ]);\n } finally {\n clock.clearTimeout(timer);\n }\n}\n",
|
|
6
6
|
"main.mjs": "import PyAI from \"@pyai/sdk\";\nimport WebSocket from \"ws\"; // Explicit transport makes this work on Node 20 too.\nimport { readFile, writeFile, mkdtemp } from \"node:fs/promises\";\nimport { join } from \"node:path\";\nimport { runSession } from \"./session.mjs\";\nimport { readWav, writeWav, assessAnswer } from \"./audio.mjs\";\n\nasync function main() {\n const [callerPath, interruptionPath, extra] = process.argv.slice(2);\n if (!callerPath || extra) throw new Error(\"Usage: node main.mjs caller.wav [interruption.wav]\");\n if (!process.env.PYAI_API_KEY) throw new Error(\"Inject PYAI_API_KEY through your environment first\");\n const caller = readWav(await readFile(callerPath));\n const interruption = interruptionPath ? readWav(await readFile(interruptionPath)) : undefined;\n const pyai = new PyAI({ apiKey: process.env.PYAI_API_KEY, maxRetries: 0,\n fetch: (url, options) => fetch(url, { ...options, signal: AbortSignal.timeout(30000) }),\n });\n const { report, audio, replyAudio } = await runSession({ pyai, webSocket: WebSocket, caller, interruption });\n const directory = await mkdtemp(join(process.cwd(), \"omni-run-\"));\n const save = (name, data) => writeFile(join(directory, name), data, { mode: 0o600, flag: \"wx\" });\n await save(\"session.wav\", writeWav(audio, report.output_rate ?? 24000));\n const reply = writeWav(replyAudio, report.output_rate ?? 24000);\n await save(\"reply.wav\", reply);\n // Hear the captured bytes, rather than trusting synthesis-advisory text.\n let capturedTranscript = \"\";\n if (report.reply_audio_bytes) {\n try {\n const heard = await pyai.audio.transcriptions.create({ file: new File([reply], \"reply.wav\", { type: \"audio/wav\" }), model: \"pyai-hear\" });\n capturedTranscript = heard.text;\n } catch { report.hear_error = \"Captured audio saved; Hear verification failed\"; }\n }\n const result = { ...report, captured_transcript: capturedTranscript, evidence: assessAnswer(report, capturedTranscript) };\n await save(\"report.json\", JSON.stringify(result, null, 2) + \"\\n\");\n console.log(JSON.stringify({ directory, ...result }, null, 2));\n if (result.evidence.captured_answer !== \"verified_by_hear\" || (interruption && result.evidence.interruption !== \"simulated_queue_cleared\")) process.exitCode = 1;\n}\n\nmain().catch(error => {\n const message = String(error.message).replaceAll(process.env.PYAI_API_KEY || \"\\0\", \"[REDACTED]\");\n console.error(JSON.stringify({ error: message, ...(error.omni_code ? { code: error.omni_code } : {}) })); process.exitCode = 1;\n});\n",
|
|
7
7
|
"session.mjs": "import { newAudioCapture, recordAudio, recordTurnBegin, streamPcmRealtime, INTERRUPTION_OBSERVATION } from \"./timing.mjs\";\n\nexport const realClock = {\n now: () => performance.now(),\n sleep: ms => new Promise(resolve => setTimeout(resolve, ms)),\n};\n\n// Only recognized machine codes may leave the example. Error messages, call\n// IDs and arbitrary code values can contain credentials or customer content.\nconst publicErrorCodes = new Set([\n \"invalid_configure\", \"unsupported_tool_transport\", \"tool_configuration_locked\",\n \"media_dead\", \"unauthorized\", \"insufficient_scope\", \"rate_limit_exceeded\",\n \"concurrency_limit_exceeded\", \"daily_cap_exceeded\", \"credit_exhausted\",\n]);\nconst transportErrorCodes = new Set([\"ECONNRESET\", \"ECONNREFUSED\", \"ETIMEDOUT\", \"ENOTFOUND\", \"EAI_AGAIN\"]);\nexport function omniFailure(error) {\n const code = error instanceof Error\n ? (transportErrorCodes.has(error.code) ? error.code : \"transport_or_protocol_error\")\n : (publicErrorCodes.has(error?.code) ? error.code : \"unrecognized_server_error\");\n return Object.assign(new Error(`Omni failed (${code}); no automatic retry was attempted`), { omni_code: code });\n}\n\n/** One bounded, server-side example. Its playback queue is simulated. */\nexport async function runSession({ pyai, webSocket, caller, interruption, clock = realClock, timeoutMs = 60000 }) {\n const rate = 24000, frameSize = 480, silence = new Int16Array(frameSize);\n if (!(caller instanceof Int16Array) || !caller.length || caller.length > rate * 20) throw new Error(\"Provide 1–20 seconds of caller PCM16 at 24 kHz\");\n if (interruption && (!(interruption instanceof Int16Array) || !interruption.length || interruption.length > rate * 20)) throw new Error(\"Invalid interruption PCM\");\n const started = clock.now(), all = newAudioCapture(started), reply = newAudioCapture(started);\n const report = { input_rate: rate, output_rate: null, playback_sink: \"simulated\", tool_calls: 0, tool_executions: 0, duplicate_tool_calls: 0, flushes_after_interruption: 0, cleared_queue_ms: 0, interruption_requested: !!interruption, caller_started_ms: null, interruption_started_ms: null, advisory_transcripts: [], end_reason: \"timeout\" };\n let session, configuredAt = null, closed = false, failure = null, phase = \"greeting\", offset = 0, queueEnd = started, interruptionSent = false, inputEndAt = started;\n const handled = new Set();\n const fail = message => { failure ??= new Error(message); };\n try {\n session = pyai.omni.connect({\n webSocket, rate,\n configure: {\n voice_id: \"stock_amos_en_us\", language: \"en\", greeting: \"Hello. How can I help?\",\n persona: \"When asked about office hours, call lookup_office_hours. Never guess. After the result, say one short sentence giving its opening_time.\",\n tools: [{ name: \"lookup_office_hours\", description: \"Read the synthetic example office opening time.\", side_effect: \"read\", parameters: { type: \"object\", properties: {}, required: [], additionalProperties: false } }],\n },\n onHello: frame => {\n const match = /^pcm16@(8000|24000)$/.exec(frame.audio_out ?? \"\");\n if (!match) return fail(\"Unsupported Omni output format\");\n report.output_rate = Number(match[1]);\n },\n onConfigured: () => { configuredAt ??= clock.now(); },\n onAudio: chunk => {\n if (!report.output_rate) return fail(\"Audio arrived before its output format\");\n const bytes = chunk instanceof ArrayBuffer ? new Uint8Array(chunk) : new Uint8Array(chunk.buffer, chunk.byteOffset, chunk.byteLength);\n if (bytes.byteLength % 2) return fail(\"Invalid PCM16 output\");\n const view = new DataView(bytes.buffer, bytes.byteOffset, bytes.byteLength);\n const pcm = Int16Array.from({ length: bytes.length / 2 }, (_, i) => view.getInt16(i * 2, true));\n const now = clock.now();\n recordAudio(all, pcm, now, report.output_rate);\n queueEnd = Math.max(now, queueEnd) + pcm.length / report.output_rate * 1000;\n if (report.tool_executions) recordAudio(reply, pcm, now, report.output_rate);\n },\n onEvent: frame => {\n if (frame.event === \"turn_begin\") recordTurnBegin(all, clock.now(), frame.turn);\n },\n onTranscript: frame => {\n // An assistant advisory is progress; never treat it as received speech.\n if (report.advisory_transcripts.length < 100) report.advisory_transcripts.push({ role: frame.role, text: frame.text });\n if (frame.role === \"assistant\") all.lastAssistantTranscriptAt = clock.now();\n },\n onToolCall: frame => {\n report.tool_calls++;\n if (handled.has(frame.call_id)) { report.duplicate_tool_calls++; return; }\n handled.add(frame.call_id);\n if (frame.name !== \"lookup_office_hours\") { session.toolResult(frame.call_id, { error: \"Unknown example tool\" }); return; }\n // Replace this read-only fixture with your authorized data lookup.\n report.tool_executions++;\n session.toolResult(frame.call_id, { result: { opening_time: \"9 a.m.\", source: \"synthetic example office\" } });\n },\n onBargeIn: () => {\n const now = clock.now(), queued = Math.max(0, queueEnd - now);\n if (report.interruption_started_ms != null) { report.flushes_after_interruption++; report.cleared_queue_ms += queued; }\n // A real speaker adapter must cancel scheduled and currently playing audio.\n queueEnd = now; all.playbackEndAt = now; reply.playbackEndAt = now;\n },\n onError: error => { failure ??= omniFailure(error); },\n onClose: () => { closed = true; },\n });\n while (clock.now() - started < timeoutMs) {\n if (failure) throw failure;\n if (closed) { report.end_reason = \"closed\"; break; }\n const now = clock.now();\n if (configuredAt == null || !report.output_rate) {\n if (now - started >= 10000) throw new Error(\"Omni connection/configuration timed out\");\n await clock.sleep(20); continue;\n }\n const quietAt = Math.max(queueEnd, all.lastPacketAt ?? now, all.lastAssistantTranscriptAt ?? 0, inputEndAt);\n if (phase === \"greeting\") {\n if (all.firstPacketAt != null && now - configuredAt >= 2000 && now >= quietAt + 800) {\n phase = \"caller\"; report.caller_started_ms = now - started;\n } else if (now - configuredAt > 15000) throw new Error(\"Greeting did not drain within 15 seconds\");\n }\n if (phase === \"reply\" && interruption && !interruptionSent && reply.samples && queueEnd - now >= 200) {\n phase = \"interruption\"; offset = 0; interruptionSent = true; report.interruption_started_ms = now - started;\n }\n const observationEnd = inputEndAt + (interruptionSent ? INTERRUPTION_OBSERVATION.minimumAfterCallerStreamMs : 0);\n if (phase === \"reply\" && reply.samples && now >= quietAt + 2000 && now >= observationEnd && (!interruption || interruptionSent)) {\n report.end_reason = \"quiet_window\"; break;\n }\n // Exactly one producer: each 20 ms slot contains caller PCM OR silence.\n let pcm = silence;\n if (phase === \"caller\" || phase === \"interruption\") {\n const source = phase === \"caller\" ? caller : interruption;\n pcm = source.subarray(offset, offset + frameSize); offset += pcm.length;\n if (offset >= source.length) {\n inputEndAt = now + pcm.length / rate * 1000;\n report[phase === \"caller\" ? \"caller_completed_ms\" : \"interruption_completed_ms\"] = inputEndAt - started;\n phase = \"reply\"; offset = 0;\n }\n }\n await streamPcmRealtime(session, pcm, rate, clock);\n }\n report.audio_bytes = all.samples * 2;\n report.reply_audio_bytes = reply.samples * 2;\n report.elapsed_ms = Math.round(clock.now() - started);\n return { report, audio: all.pcm, replyAudio: reply.pcm };\n } finally { session?.close(); }\n}\n",
|
package/src/dialer/index.ts
CHANGED
|
@@ -2,6 +2,7 @@ export * from "./managedDialer.ts";
|
|
|
2
2
|
export * from "./dialingQueue.ts";
|
|
3
3
|
export * from "./indexedDbJournal.ts";
|
|
4
4
|
export * from "./voiceSdkAudio.ts";
|
|
5
|
+
export * from "./managedBrowserAudio.ts";
|
|
5
6
|
export * from "./meetingRecorder.ts";
|
|
6
7
|
export * from "./meetingRecordingStore.ts";
|
|
7
8
|
|
|
@@ -0,0 +1,75 @@
|
|
|
1
|
+
import type { DialerAudio } from "./managedDialer.ts";
|
|
2
|
+
import type { TelephonyBrowserSession } from "../index.ts";
|
|
3
|
+
import { VoiceSdkAudio } from "./voiceSdkAudio.ts";
|
|
4
|
+
|
|
5
|
+
/** Issued by the application's server after PyAI has authorized one call.
|
|
6
|
+
* This contains no workspace key or reusable carrier credential. */
|
|
7
|
+
export type BrowserAudioCapability = Pick<TelephonyBrowserSession, "access_token" | "connect_params"> & {
|
|
8
|
+
transport?: "voice-sdk";
|
|
9
|
+
};
|
|
10
|
+
type AudioEvent = Parameters<DialerAudio["connect"]>[1];
|
|
11
|
+
export interface ManagedBrowserAudio {
|
|
12
|
+
connect(capability: BrowserAudioCapability, event: AudioEvent): Promise<void>;
|
|
13
|
+
disconnect(): void;
|
|
14
|
+
mute(value: boolean): void;
|
|
15
|
+
sendDigits(digits: string): void;
|
|
16
|
+
}
|
|
17
|
+
export interface ManagedBrowserAudioOptions {
|
|
18
|
+
/** Optional media adapter for embedded hosts. Ordinary applications omit it. */
|
|
19
|
+
loadTransport?: (transport: "voice-sdk") => Promise<DialerAudio>;
|
|
20
|
+
}
|
|
21
|
+
|
|
22
|
+
async function loadTransport(): Promise<DialerAudio> {
|
|
23
|
+
// The upstream transport is an implementation detail of this PyAI adapter.
|
|
24
|
+
// Applications must not import or construct the carrier's Device themselves.
|
|
25
|
+
const { Device } = await import("@twilio/voice-sdk");
|
|
26
|
+
type DeviceShape = ReturnType<ConstructorParameters<typeof VoiceSdkAudio>[0]>;
|
|
27
|
+
return new VoiceSdkAudio(token => new Device(token) as unknown as DeviceShape);
|
|
28
|
+
}
|
|
29
|
+
|
|
30
|
+
/** Provider-owned, lazy media setup. A cancelled import cannot place a late call. */
|
|
31
|
+
export function createManagedBrowserAudio(options: ManagedBrowserAudioOptions = {}): ManagedBrowserAudio {
|
|
32
|
+
let generation = 0;
|
|
33
|
+
let audio: DialerAudio | null = null;
|
|
34
|
+
const disconnect = () => {
|
|
35
|
+
generation++;
|
|
36
|
+
const prior = audio; audio = null;
|
|
37
|
+
try { prior?.disconnect(); } catch { /* Remote call cleanup belongs to the server broker. */ }
|
|
38
|
+
};
|
|
39
|
+
return {
|
|
40
|
+
async connect(capability, event) {
|
|
41
|
+
const params = capability.connect_params;
|
|
42
|
+
if (!capability.access_token || capability.transport !== undefined && capability.transport !== "voice-sdk" ||
|
|
43
|
+
!params || Object.keys(params).length !== 1 || typeof params.session_id !== "string" ||
|
|
44
|
+
params.session_id.length < 1 || params.session_id.length > 200) {
|
|
45
|
+
throw new Error("A supported, single-call PyAI audio capability is required.");
|
|
46
|
+
}
|
|
47
|
+
// Capture before awaiting a lazy import; never retain a mutable caller object.
|
|
48
|
+
const session = { access_token: capability.access_token, connect_params: { session_id: params.session_id } };
|
|
49
|
+
disconnect();
|
|
50
|
+
const current = generation;
|
|
51
|
+
const candidate = await (options.loadTransport ?? loadTransport)("voice-sdk");
|
|
52
|
+
if (current !== generation) { try { candidate.disconnect(); } catch { /* Not connected. */ } return; }
|
|
53
|
+
audio = candidate;
|
|
54
|
+
try {
|
|
55
|
+
// DialerAudio implementations consume this same opaque media capability.
|
|
56
|
+
await candidate.connect(session as TelephonyBrowserSession, state => {
|
|
57
|
+
if (current === generation) event(state);
|
|
58
|
+
});
|
|
59
|
+
} catch (error) {
|
|
60
|
+
if (current === generation) disconnect();
|
|
61
|
+
throw error;
|
|
62
|
+
}
|
|
63
|
+
},
|
|
64
|
+
disconnect,
|
|
65
|
+
mute(value) {
|
|
66
|
+
if (!audio) throw new Error("No active audio.");
|
|
67
|
+
audio.mute(value);
|
|
68
|
+
},
|
|
69
|
+
sendDigits(digits) {
|
|
70
|
+
if (!audio) throw new Error("No active audio.");
|
|
71
|
+
if (!/^[0-9*#]{1,32}$/.test(digits)) throw new Error("Use one to 32 keypad digits.");
|
|
72
|
+
audio.sendDigits(digits);
|
|
73
|
+
},
|
|
74
|
+
};
|
|
75
|
+
}
|
|
@@ -6,6 +6,7 @@ export interface DialerBroker {
|
|
|
6
6
|
create(input: TelephonyBrowserSessionCreate, idempotencyKey: string): Promise<TelephonyBrowserSession>;
|
|
7
7
|
get(callId: string): Promise<TelephonyCall>;
|
|
8
8
|
hangup(callId: string): Promise<TelephonyCall>;
|
|
9
|
+
setHold?(callId: string, held: boolean, key: string): Promise<TelephonyCall>;
|
|
9
10
|
transfers?: {
|
|
10
11
|
create(callId: string, input: { to: string; mode: "warm" | "cold" }, key: string): Promise<TelephonyCall>;
|
|
11
12
|
complete(callId: string, transferId: string): Promise<TelephonyCall>;
|
|
@@ -24,6 +24,12 @@ export function createServerDialerBroker(sdk: PyAI, authenticatedUserId: string)
|
|
|
24
24
|
if (dialerCallHandedOff(call)) return call;
|
|
25
25
|
return sdk.telephony.calls.hangup(id);
|
|
26
26
|
},
|
|
27
|
+
async setHold(id, held, key) {
|
|
28
|
+
await ownedCall(id);
|
|
29
|
+
if (!key || key.length > 255 || /[\r\n]/.test(key)) throw new Error("A valid hold retry key is required.");
|
|
30
|
+
const scopedKey = "dialer-hold:" + createHash("sha256").update(authenticatedUserId).update("\0").update(id).update("\0").update(key).digest("hex");
|
|
31
|
+
return sdk.telephony.calls.setHold(id, { held, user_id: authenticatedUserId }, { idempotencyKey: scopedKey });
|
|
32
|
+
},
|
|
27
33
|
transfers: {
|
|
28
34
|
async create(id, input, key) {
|
|
29
35
|
await ownedCall(id);
|
package/src/index.ts
CHANGED
|
@@ -1501,6 +1501,8 @@ export interface TelephonyCall {
|
|
|
1501
1501
|
agent_id: string | null;
|
|
1502
1502
|
call_type?: "human";
|
|
1503
1503
|
transfer_enabled?: boolean;
|
|
1504
|
+
held?: boolean;
|
|
1505
|
+
hold_pending?: boolean;
|
|
1504
1506
|
transfers?: TelephonyTransfer[];
|
|
1505
1507
|
user_id?: string;
|
|
1506
1508
|
client_reference: string | null;
|
|
@@ -1518,6 +1520,32 @@ export interface TelephonyTransfer {
|
|
|
1518
1520
|
status: "holding" | "dialing" | "ready" | "completing" | "canceling" | "completed" | "canceled" | "failed";
|
|
1519
1521
|
to: string | null;
|
|
1520
1522
|
}
|
|
1523
|
+
export interface TelephonyCallQuality {
|
|
1524
|
+
call_id: string;
|
|
1525
|
+
status: "available" | "unavailable";
|
|
1526
|
+
reason: string | null;
|
|
1527
|
+
source: "managed_transport" | null;
|
|
1528
|
+
scope: "caller_to_platform" | null;
|
|
1529
|
+
observation: "terminal_snapshot" | "partial_snapshot" | null;
|
|
1530
|
+
observed_at: string | null;
|
|
1531
|
+
live: false;
|
|
1532
|
+
metrics: Record<"packets_received" | "packets_lost" | "packets_reordered" | "duplicate_packets" |
|
|
1533
|
+
"packets_sent" | "playout_frames" | "concealed_frames" | "late_frames" | "playout_duplicates" |
|
|
1534
|
+
"jitter_ms" | "packet_loss_percent" | "concealment_percent", number | null> & { mos: null; rtt_ms: null };
|
|
1535
|
+
unavailable_metrics: string[];
|
|
1536
|
+
}
|
|
1537
|
+
export interface TelephonyCallCapabilities {
|
|
1538
|
+
object: "telephony.call_capabilities";
|
|
1539
|
+
call_id: string;
|
|
1540
|
+
version: 1;
|
|
1541
|
+
call_type: "human" | "ai";
|
|
1542
|
+
controls: Record<"hangup" | "hold" | "warm_transfer" | "cold_transfer" | "complete_transfer" |
|
|
1543
|
+
"cancel_transfer" | "listen" | "whisper" | "barge", { supported: boolean; reason: string | null }>;
|
|
1544
|
+
client_controls: Array<"mute" | "dtmf">;
|
|
1545
|
+
quality: { read_supported: true; live: false };
|
|
1546
|
+
revalidation_required: true;
|
|
1547
|
+
}
|
|
1548
|
+
|
|
1521
1549
|
export interface TelephonyCallCreate {
|
|
1522
1550
|
from_number_id: string;
|
|
1523
1551
|
to: string;
|
|
@@ -2396,6 +2424,8 @@ export class PyAI {
|
|
|
2396
2424
|
create: (input: TelephonyCallCreate, options: { idempotencyKey: string }): Promise<TelephonyCall> =>
|
|
2397
2425
|
this.postJson("/v1/telephony/calls", input, telephonyIdempotencyHeader(options.idempotencyKey)),
|
|
2398
2426
|
get: (id: string): Promise<TelephonyCall> => this.getJson(`/v1/telephony/calls/${encodeURIComponent(id)}`),
|
|
2427
|
+
capabilities: (id: string): Promise<TelephonyCallCapabilities> => this.getJson(`/v1/telephony/calls/${encodeURIComponent(id)}/capabilities`),
|
|
2428
|
+
quality: (id: string): Promise<TelephonyCallQuality> => this.getJson(`/v1/telephony/calls/${encodeURIComponent(id)}/quality`),
|
|
2399
2429
|
list: (input: { limit?: number; cursor?: string; direction?: "inbound" | "outbound"; status?: TelephonyCallStatus } = {}): Promise<ListResponse<TelephonyCall> & { next_cursor: string | null }> => {
|
|
2400
2430
|
const query = new URLSearchParams();
|
|
2401
2431
|
for (const [key, value] of Object.entries(input)) if (value !== undefined) query.set(key, String(value));
|
|
@@ -2408,6 +2438,9 @@ export class PyAI {
|
|
|
2408
2438
|
recording: (id: string, options: { range?: string } = {}): Promise<Response> =>
|
|
2409
2439
|
this.request(`/v1/telephony/calls/${encodeURIComponent(id)}/recording`, { headers: this.authHeaders(options.range ? { Range: options.range } : {}) }),
|
|
2410
2440
|
deleteRecording: async (id: string): Promise<void> => { await this.deleteReq(`/v1/telephony/calls/${encodeURIComponent(id)}/recording`); },
|
|
2441
|
+
/** The same durable key recovers an uncertain hold command without repeating it. */
|
|
2442
|
+
setHold: (id: string, input: { user_id: string; held: boolean }, options: { idempotencyKey: string }): Promise<TelephonyCall> =>
|
|
2443
|
+
this.postJson(`/v1/telephony/calls/${encodeURIComponent(id)}/hold`, input, telephonyIdempotencyHeader(options.idempotencyKey)),
|
|
2411
2444
|
transfers: {
|
|
2412
2445
|
/** Persist the key and target before requesting a consultation. */
|
|
2413
2446
|
create: (id: string, input: { to: string; mode: "warm" | "cold"; user_id: string }, options: { idempotencyKey: string }): Promise<TelephonyCall> =>
|