@jambonz/sdk 0.1.31 → 0.2.0
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/dist/client/index.cjs +6 -0
- package/dist/client/index.cjs.map +1 -1
- package/dist/client/index.d.cts +2 -2
- package/dist/client/index.d.ts +2 -2
- package/dist/client/index.js +6 -0
- package/dist/client/index.js.map +1 -1
- package/dist/index.cjs +41 -85
- package/dist/index.cjs.map +1 -1
- package/dist/index.d.cts +3 -3
- package/dist/index.d.ts +3 -3
- package/dist/index.js +43 -87
- package/dist/index.js.map +1 -1
- package/dist/{rest--MT5-UzW.d.cts → rest-B_GQITxM.d.cts} +1 -1
- package/dist/{rest-By6fi3ZV.d.ts → rest-DgctQy2b.d.ts} +1 -1
- package/dist/types/index.d.cts +2 -2
- package/dist/types/index.d.ts +2 -2
- package/dist/{verb-builder-CnqpQcT3.d.ts → verb-builder-B6zPuGd1.d.cts} +2 -2
- package/dist/{verb-builder-DHzc_LPM.d.cts → verb-builder-B7RIVi0t.d.ts} +2 -2
- package/dist/{verbs-DB-J_0eQ.d.cts → verbs-B18Kazd9.d.cts} +2 -0
- package/dist/{verbs-DB-J_0eQ.d.ts → verbs-B18Kazd9.d.ts} +2 -0
- package/dist/webhook/index.cjs +35 -85
- package/dist/webhook/index.cjs.map +1 -1
- package/dist/webhook/index.d.cts +2 -2
- package/dist/webhook/index.d.ts +2 -2
- package/dist/webhook/index.js +37 -87
- package/dist/webhook/index.js.map +1 -1
- package/dist/websocket/index.cjs +35 -85
- package/dist/websocket/index.cjs.map +1 -1
- package/dist/websocket/index.d.cts +2 -2
- package/dist/websocket/index.d.ts +2 -2
- package/dist/websocket/index.js +37 -87
- package/dist/websocket/index.js.map +1 -1
- package/package.json +4 -4
- package/schema/callbacks/amd.schema.json +0 -50
- package/schema/callbacks/base.schema.json +0 -29
- package/schema/callbacks/call-status.schema.json +0 -22
- package/schema/callbacks/conference-status.schema.json +0 -24
- package/schema/callbacks/conference-wait.schema.json +0 -11
- package/schema/callbacks/conference.schema.json +0 -11
- package/schema/callbacks/dequeue.schema.json +0 -19
- package/schema/callbacks/dial-dtmf.schema.json +0 -18
- package/schema/callbacks/dial-hold.schema.json +0 -22
- package/schema/callbacks/dial-refer.schema.json +0 -28
- package/schema/callbacks/dial.schema.json +0 -31
- package/schema/callbacks/enqueue-wait.schema.json +0 -17
- package/schema/callbacks/enqueue.schema.json +0 -27
- package/schema/callbacks/gather-partial.schema.json +0 -54
- package/schema/callbacks/gather.schema.json +0 -60
- package/schema/callbacks/listen.schema.json +0 -21
- package/schema/callbacks/llm.schema.json +0 -30
- package/schema/callbacks/message.schema.json +0 -35
- package/schema/callbacks/pipeline-turn.schema.json +0 -109
- package/schema/callbacks/play.schema.json +0 -36
- package/schema/callbacks/session-new.schema.json +0 -143
- package/schema/callbacks/session-reconnect.schema.json +0 -9
- package/schema/callbacks/session-redirect.schema.json +0 -38
- package/schema/callbacks/sip-refer-event.schema.json +0 -20
- package/schema/callbacks/sip-refer.schema.json +0 -22
- package/schema/callbacks/sip-request.schema.json +0 -27
- package/schema/callbacks/transcribe-translation.schema.json +0 -24
- package/schema/callbacks/transcribe.schema.json +0 -46
- package/schema/callbacks/tts-streaming-event.schema.json +0 -77
- package/schema/callbacks/verb-status.schema.json +0 -57
- package/schema/components/actionHook.schema.json +0 -36
- package/schema/components/actionHookDelayAction.schema.json +0 -37
- package/schema/components/amd.schema.json +0 -68
- package/schema/components/auth.schema.json +0 -18
- package/schema/components/bidirectionalAudio.schema.json +0 -22
- package/schema/components/fillerNoise.schema.json +0 -25
- package/schema/components/llm-base.schema.json +0 -94
- package/schema/components/recognizer-assemblyAiOptions.schema.json +0 -66
- package/schema/components/recognizer-awsOptions.schema.json +0 -52
- package/schema/components/recognizer-azureOptions.schema.json +0 -32
- package/schema/components/recognizer-cobaltOptions.schema.json +0 -34
- package/schema/components/recognizer-customOptions.schema.json +0 -27
- package/schema/components/recognizer-deepgramOptions.schema.json +0 -147
- package/schema/components/recognizer-elevenlabsOptions.schema.json +0 -39
- package/schema/components/recognizer-gladiaOptions.schema.json +0 -8
- package/schema/components/recognizer-googleOptions.schema.json +0 -35
- package/schema/components/recognizer-houndifyOptions.schema.json +0 -53
- package/schema/components/recognizer-ibmOptions.schema.json +0 -54
- package/schema/components/recognizer-nuanceOptions.schema.json +0 -150
- package/schema/components/recognizer-nvidiaOptions.schema.json +0 -39
- package/schema/components/recognizer-openaiOptions.schema.json +0 -59
- package/schema/components/recognizer-sonioxOptions.schema.json +0 -46
- package/schema/components/recognizer-speechmaticsOptions.schema.json +0 -100
- package/schema/components/recognizer-verbioOptions.schema.json +0 -46
- package/schema/components/recognizer.schema.json +0 -216
- package/schema/components/synthesizer.schema.json +0 -82
- package/schema/components/target.schema.json +0 -105
- package/schema/components/vad.schema.json +0 -48
- package/schema/jambonz-app.schema.json +0 -112
- package/schema/verbs/alert.schema.json +0 -34
- package/schema/verbs/answer.schema.json +0 -22
- package/schema/verbs/conference.schema.json +0 -107
- package/schema/verbs/config.schema.json +0 -218
- package/schema/verbs/deepgram_s2s.schema.json +0 -81
- package/schema/verbs/dequeue.schema.json +0 -51
- package/schema/verbs/dial.schema.json +0 -187
- package/schema/verbs/dialogflow.schema.json +0 -148
- package/schema/verbs/dtmf.schema.json +0 -49
- package/schema/verbs/dub.schema.json +0 -103
- package/schema/verbs/elevenlabs_s2s.schema.json +0 -81
- package/schema/verbs/enqueue.schema.json +0 -53
- package/schema/verbs/gather.schema.json +0 -188
- package/schema/verbs/google_s2s.schema.json +0 -42
- package/schema/verbs/hangup.schema.json +0 -36
- package/schema/verbs/leave.schema.json +0 -22
- package/schema/verbs/listen.schema.json +0 -127
- package/schema/verbs/llm.schema.json +0 -44
- package/schema/verbs/message.schema.json +0 -82
- package/schema/verbs/openai_s2s.schema.json +0 -42
- package/schema/verbs/pause.schema.json +0 -36
- package/schema/verbs/pipeline.schema.json +0 -234
- package/schema/verbs/play.schema.json +0 -96
- package/schema/verbs/redirect.schema.json +0 -34
- package/schema/verbs/s2s.schema.json +0 -39
- package/schema/verbs/say.schema.json +0 -107
- package/schema/verbs/sip-decline.schema.json +0 -58
- package/schema/verbs/sip-refer.schema.json +0 -58
- package/schema/verbs/sip-request.schema.json +0 -54
- package/schema/verbs/stream.schema.json +0 -103
- package/schema/verbs/tag.schema.json +0 -41
- package/schema/verbs/transcribe.schema.json +0 -57
- package/schema/verbs/ultravox_s2s.schema.json +0 -41
|
@@ -1,36 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/hangup",
|
|
4
|
-
"minVersion": "0.9.6",
|
|
5
|
-
"title": "Hangup",
|
|
6
|
-
"description": "Terminates the call. Optionally includes custom SIP headers on the BYE request.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"properties": {
|
|
9
|
-
"verb": {
|
|
10
|
-
"const": "hangup",
|
|
11
|
-
"description": "The verb name."
|
|
12
|
-
},
|
|
13
|
-
"id": {
|
|
14
|
-
"type": "string",
|
|
15
|
-
"description": "An optional unique identifier for this verb instance."
|
|
16
|
-
},
|
|
17
|
-
"headers": {
|
|
18
|
-
"type": "object",
|
|
19
|
-
"description": "Custom SIP headers to include on the BYE request.",
|
|
20
|
-
"additionalProperties": {
|
|
21
|
-
"type": "string"
|
|
22
|
-
}
|
|
23
|
-
}
|
|
24
|
-
},
|
|
25
|
-
"examples": [
|
|
26
|
-
{
|
|
27
|
-
"verb": "hangup"
|
|
28
|
-
},
|
|
29
|
-
{
|
|
30
|
-
"verb": "hangup",
|
|
31
|
-
"headers": {
|
|
32
|
-
"X-Reason": "call-complete"
|
|
33
|
-
}
|
|
34
|
-
}
|
|
35
|
-
]
|
|
36
|
-
}
|
|
@@ -1,22 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/leave",
|
|
4
|
-
"minVersion": "0.9.6",
|
|
5
|
-
"title": "Leave",
|
|
6
|
-
"description": "Removes the caller from a conference or queue that they are currently in. Execution continues with the next verb in the application.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"properties": {
|
|
9
|
-
"verb": {
|
|
10
|
-
"const": "leave"
|
|
11
|
-
},
|
|
12
|
-
"id": {
|
|
13
|
-
"type": "string",
|
|
14
|
-
"description": "An optional unique identifier for this verb instance."
|
|
15
|
-
}
|
|
16
|
-
},
|
|
17
|
-
"examples": [
|
|
18
|
-
{
|
|
19
|
-
"verb": "leave"
|
|
20
|
-
}
|
|
21
|
-
]
|
|
22
|
-
}
|
|
@@ -1,127 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/listen",
|
|
4
|
-
"minVersion": "0.9.6",
|
|
5
|
-
"title": "Listen",
|
|
6
|
-
"description": "Streams real-time call audio to an external websocket endpoint. The remote endpoint receives raw audio and can optionally send audio back (bidirectional). Used for custom speech processing, real-time analysis, AI agent integration, and recording to external systems.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"properties": {
|
|
9
|
-
"verb": {
|
|
10
|
-
"const": "listen",
|
|
11
|
-
"description": "The verb name."
|
|
12
|
-
},
|
|
13
|
-
"id": {
|
|
14
|
-
"type": "string",
|
|
15
|
-
"description": "An optional unique identifier for this verb instance."
|
|
16
|
-
},
|
|
17
|
-
"url": {
|
|
18
|
-
"type": "string",
|
|
19
|
-
"format": "uri",
|
|
20
|
-
"description": "The websocket URL to stream audio to.",
|
|
21
|
-
"examples": [
|
|
22
|
-
"wss://myapp.example.com/audio-stream"
|
|
23
|
-
]
|
|
24
|
-
},
|
|
25
|
-
"actionHook": {
|
|
26
|
-
"$ref": "../components/actionHook",
|
|
27
|
-
"description": "A webhook invoked when the listen session ends. Should return the next verbs to execute."
|
|
28
|
-
},
|
|
29
|
-
"wsAuth": {
|
|
30
|
-
"$ref": "../components/auth",
|
|
31
|
-
"description": "Authentication credentials for the websocket connection."
|
|
32
|
-
},
|
|
33
|
-
"mixType": {
|
|
34
|
-
"type": "string",
|
|
35
|
-
"description": "How to mix the audio channels when streaming. 'mono' sends a single mixed channel, 'stereo' sends caller and callee as separate left/right channels, 'mixed' sends both as a single mixed stream.",
|
|
36
|
-
"enum": [
|
|
37
|
-
"mono",
|
|
38
|
-
"stereo",
|
|
39
|
-
"mixed"
|
|
40
|
-
],
|
|
41
|
-
"default": "mono"
|
|
42
|
-
},
|
|
43
|
-
"metadata": {
|
|
44
|
-
"type": "object",
|
|
45
|
-
"description": "Arbitrary metadata to send to the websocket endpoint in the initial connection message.",
|
|
46
|
-
"additionalProperties": true
|
|
47
|
-
},
|
|
48
|
-
"sampleRate": {
|
|
49
|
-
"type": "number",
|
|
50
|
-
"description": "The audio sample rate in Hz.",
|
|
51
|
-
"examples": [
|
|
52
|
-
8000,
|
|
53
|
-
16000,
|
|
54
|
-
24000
|
|
55
|
-
],
|
|
56
|
-
"default": 8000
|
|
57
|
-
},
|
|
58
|
-
"finishOnKey": {
|
|
59
|
-
"type": "string",
|
|
60
|
-
"description": "A DTMF key that ends the listen session when pressed.",
|
|
61
|
-
"examples": [
|
|
62
|
-
"#"
|
|
63
|
-
]
|
|
64
|
-
},
|
|
65
|
-
"maxLength": {
|
|
66
|
-
"type": "number",
|
|
67
|
-
"description": "Maximum duration in seconds for the listen session."
|
|
68
|
-
},
|
|
69
|
-
"passDtmf": {
|
|
70
|
-
"type": "boolean",
|
|
71
|
-
"description": "If true, forward DTMF events to the websocket endpoint."
|
|
72
|
-
},
|
|
73
|
-
"playBeep": {
|
|
74
|
-
"type": "boolean",
|
|
75
|
-
"description": "If true, play a beep tone before streaming begins."
|
|
76
|
-
},
|
|
77
|
-
"disableBidirectionalAudio": {
|
|
78
|
-
"type": "boolean",
|
|
79
|
-
"description": "If true, disable receiving audio from the websocket endpoint. Audio flows only from the call to the websocket, not back."
|
|
80
|
-
},
|
|
81
|
-
"bidirectionalAudio": {
|
|
82
|
-
"$ref": "../components/bidirectionalAudio",
|
|
83
|
-
"description": "Fine-grained configuration for bidirectional audio."
|
|
84
|
-
},
|
|
85
|
-
"timeout": {
|
|
86
|
-
"type": "number",
|
|
87
|
-
"description": "Time in seconds to wait for audio activity before ending the listen session."
|
|
88
|
-
},
|
|
89
|
-
"transcribe": {
|
|
90
|
-
"$ref": "transcribe",
|
|
91
|
-
"description": "Nested transcribe verb — enables simultaneous real-time transcription of the audio being streamed."
|
|
92
|
-
},
|
|
93
|
-
"earlyMedia": {
|
|
94
|
-
"type": "boolean",
|
|
95
|
-
"description": "If true, begin streaming audio before the call is formally answered."
|
|
96
|
-
},
|
|
97
|
-
"channel": {
|
|
98
|
-
"type": "number",
|
|
99
|
-
"description": "Specific audio channel to stream. Used when streaming a single channel of a multi-channel call."
|
|
100
|
-
}
|
|
101
|
-
},
|
|
102
|
-
"required": [
|
|
103
|
-
"url"
|
|
104
|
-
],
|
|
105
|
-
"examples": [
|
|
106
|
-
{
|
|
107
|
-
"verb": "listen",
|
|
108
|
-
"url": "wss://myapp.example.com/audio-stream",
|
|
109
|
-
"actionHook": "/listen-complete",
|
|
110
|
-
"sampleRate": 16000,
|
|
111
|
-
"mixType": "stereo"
|
|
112
|
-
},
|
|
113
|
-
{
|
|
114
|
-
"verb": "listen",
|
|
115
|
-
"url": "wss://myapp.example.com/ai-agent",
|
|
116
|
-
"bidirectionalAudio": {
|
|
117
|
-
"enabled": true,
|
|
118
|
-
"streaming": true,
|
|
119
|
-
"sampleRate": 24000
|
|
120
|
-
},
|
|
121
|
-
"metadata": {
|
|
122
|
-
"callType": "support",
|
|
123
|
-
"language": "en-US"
|
|
124
|
-
}
|
|
125
|
-
}
|
|
126
|
-
]
|
|
127
|
-
}
|
|
@@ -1,44 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/llm",
|
|
4
|
-
"minVersion": "0.9.6",
|
|
5
|
-
"title": "LLM",
|
|
6
|
-
"description": "Connects the caller to a large language model for a real-time voice conversation. Handles the complete STT → LLM → TTS pipeline, including turn detection, interruption handling, and tool/function calling. The caller speaks naturally and the LLM responds via synthesized speech. This is the primary verb for building AI voice agents on jambonz.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"allOf": [
|
|
9
|
-
{
|
|
10
|
-
"$ref": "../components/llm-base"
|
|
11
|
-
}
|
|
12
|
-
],
|
|
13
|
-
"properties": {
|
|
14
|
-
"verb": {
|
|
15
|
-
"const": "llm",
|
|
16
|
-
"description": "The verb name."
|
|
17
|
-
}
|
|
18
|
-
},
|
|
19
|
-
"required": [
|
|
20
|
-
"vendor",
|
|
21
|
-
"llmOptions"
|
|
22
|
-
],
|
|
23
|
-
"examples": [
|
|
24
|
-
{
|
|
25
|
-
"verb": "llm",
|
|
26
|
-
"vendor": "openai",
|
|
27
|
-
"model": "gpt-4o",
|
|
28
|
-
"auth": {
|
|
29
|
-
"apiKey": "sk-..."
|
|
30
|
-
},
|
|
31
|
-
"llmOptions": {
|
|
32
|
-
"messages": [
|
|
33
|
-
{
|
|
34
|
-
"role": "system",
|
|
35
|
-
"content": "You are a helpful customer service agent. Be concise and friendly."
|
|
36
|
-
}
|
|
37
|
-
],
|
|
38
|
-
"temperature": 0.7
|
|
39
|
-
},
|
|
40
|
-
"actionHook": "/llm-complete",
|
|
41
|
-
"toolHook": "/llm-tool-call"
|
|
42
|
-
}
|
|
43
|
-
]
|
|
44
|
-
}
|
|
@@ -1,82 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/message",
|
|
4
|
-
"minVersion": "0.9.6",
|
|
5
|
-
"title": "Message",
|
|
6
|
-
"description": "Sends an SMS or MMS message. Can be used during a voice call to send a text message to the caller or another party, or as a standalone action.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"properties": {
|
|
9
|
-
"verb": {
|
|
10
|
-
"const": "message"
|
|
11
|
-
},
|
|
12
|
-
"id": {
|
|
13
|
-
"type": "string",
|
|
14
|
-
"description": "An optional unique identifier for this verb instance."
|
|
15
|
-
},
|
|
16
|
-
"to": {
|
|
17
|
-
"type": "string",
|
|
18
|
-
"description": "The destination phone number in E.164 format.",
|
|
19
|
-
"examples": [
|
|
20
|
-
"+15085551212"
|
|
21
|
-
]
|
|
22
|
-
},
|
|
23
|
-
"from": {
|
|
24
|
-
"type": "string",
|
|
25
|
-
"description": "The sender phone number in E.164 format. Must be a number provisioned on the jambonz platform.",
|
|
26
|
-
"examples": [
|
|
27
|
-
"+15085559876"
|
|
28
|
-
]
|
|
29
|
-
},
|
|
30
|
-
"text": {
|
|
31
|
-
"type": "string",
|
|
32
|
-
"description": "The text content of the message."
|
|
33
|
-
},
|
|
34
|
-
"media": {
|
|
35
|
-
"oneOf": [
|
|
36
|
-
{
|
|
37
|
-
"type": "string",
|
|
38
|
-
"format": "uri"
|
|
39
|
-
},
|
|
40
|
-
{
|
|
41
|
-
"type": "array",
|
|
42
|
-
"items": {
|
|
43
|
-
"type": "string",
|
|
44
|
-
"format": "uri"
|
|
45
|
-
}
|
|
46
|
-
}
|
|
47
|
-
],
|
|
48
|
-
"description": "URL(s) of media to attach to the message (MMS). Can be images, audio, or video.",
|
|
49
|
-
"examples": [
|
|
50
|
-
"https://example.com/images/receipt.png"
|
|
51
|
-
]
|
|
52
|
-
},
|
|
53
|
-
"carrier": {
|
|
54
|
-
"type": "string",
|
|
55
|
-
"description": "The messaging carrier to use. If not specified, the default carrier is used."
|
|
56
|
-
},
|
|
57
|
-
"account_sid": {
|
|
58
|
-
"type": "string",
|
|
59
|
-
"description": "The account SID to use for sending. Defaults to the current account."
|
|
60
|
-
},
|
|
61
|
-
"message_sid": {
|
|
62
|
-
"type": "string",
|
|
63
|
-
"description": "An optional message SID for tracking."
|
|
64
|
-
},
|
|
65
|
-
"actionHook": {
|
|
66
|
-
"$ref": "../components/actionHook",
|
|
67
|
-
"description": "A webhook invoked when the message send completes or fails."
|
|
68
|
-
}
|
|
69
|
-
},
|
|
70
|
-
"required": [
|
|
71
|
-
"to",
|
|
72
|
-
"from"
|
|
73
|
-
],
|
|
74
|
-
"examples": [
|
|
75
|
-
{
|
|
76
|
-
"verb": "message",
|
|
77
|
-
"to": "+15085551212",
|
|
78
|
-
"from": "+15085559876",
|
|
79
|
-
"text": "Your order has been confirmed. Order #12345."
|
|
80
|
-
}
|
|
81
|
-
]
|
|
82
|
-
}
|
|
@@ -1,42 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/openai_s2s",
|
|
4
|
-
"minVersion": "10.1.0",
|
|
5
|
-
"title": "OpenAI S2S",
|
|
6
|
-
"description": "Shortcut for 'llm' with vendor automatically set to 'openai'. Connects the caller to an OpenAI model for real-time speech-to-speech voice conversation.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"allOf": [
|
|
9
|
-
{
|
|
10
|
-
"$ref": "../components/llm-base"
|
|
11
|
-
}
|
|
12
|
-
],
|
|
13
|
-
"properties": {
|
|
14
|
-
"verb": {
|
|
15
|
-
"const": "openai_s2s",
|
|
16
|
-
"description": "The verb name."
|
|
17
|
-
},
|
|
18
|
-
"vendor": {
|
|
19
|
-
"type": "string",
|
|
20
|
-
"const": "openai",
|
|
21
|
-
"description": "The LLM vendor (always 'openai' for this shortcut)."
|
|
22
|
-
}
|
|
23
|
-
},
|
|
24
|
-
"required": [
|
|
25
|
-
"llmOptions"
|
|
26
|
-
],
|
|
27
|
-
"examples": [
|
|
28
|
-
{
|
|
29
|
-
"verb": "openai_s2s",
|
|
30
|
-
"model": "gpt-4o-realtime",
|
|
31
|
-
"llmOptions": {
|
|
32
|
-
"messages": [
|
|
33
|
-
{
|
|
34
|
-
"role": "system",
|
|
35
|
-
"content": "You are a helpful voice assistant."
|
|
36
|
-
}
|
|
37
|
-
]
|
|
38
|
-
},
|
|
39
|
-
"actionHook": "/s2s-complete"
|
|
40
|
-
}
|
|
41
|
-
]
|
|
42
|
-
}
|
|
@@ -1,36 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/pause",
|
|
4
|
-
"minVersion": "0.9.6",
|
|
5
|
-
"title": "Pause",
|
|
6
|
-
"description": "Pauses execution for a specified number of seconds. The caller hears silence during the pause. Useful for adding delays between verbs.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"properties": {
|
|
9
|
-
"verb": {
|
|
10
|
-
"const": "pause",
|
|
11
|
-
"description": "The verb name."
|
|
12
|
-
},
|
|
13
|
-
"id": {
|
|
14
|
-
"type": "string",
|
|
15
|
-
"description": "An optional unique identifier for this verb instance."
|
|
16
|
-
},
|
|
17
|
-
"length": {
|
|
18
|
-
"type": "number",
|
|
19
|
-
"description": "The duration of the pause in seconds.",
|
|
20
|
-
"examples": [
|
|
21
|
-
1,
|
|
22
|
-
2,
|
|
23
|
-
5
|
|
24
|
-
]
|
|
25
|
-
}
|
|
26
|
-
},
|
|
27
|
-
"required": [
|
|
28
|
-
"length"
|
|
29
|
-
],
|
|
30
|
-
"examples": [
|
|
31
|
-
{
|
|
32
|
-
"verb": "pause",
|
|
33
|
-
"length": 2
|
|
34
|
-
}
|
|
35
|
-
]
|
|
36
|
-
}
|
|
@@ -1,234 +0,0 @@
|
|
|
1
|
-
{
|
|
2
|
-
"$schema": "https://json-schema.org/draft/2020-12/schema",
|
|
3
|
-
"$id": "https://jambonz.org/schema/verbs/pipeline",
|
|
4
|
-
"minVersion": "10.1.0",
|
|
5
|
-
"title": "Pipeline",
|
|
6
|
-
"description": "Configures a complete STT → LLM → TTS voice AI pipeline with integrated turn detection. Provides a higher-level abstraction than manually orchestrating the individual components. Optimized for building voice AI agents with proper turn-taking behavior.",
|
|
7
|
-
"type": "object",
|
|
8
|
-
"properties": {
|
|
9
|
-
"verb": {
|
|
10
|
-
"const": "pipeline"
|
|
11
|
-
},
|
|
12
|
-
"id": {
|
|
13
|
-
"type": "string",
|
|
14
|
-
"description": "An optional unique identifier for this verb instance."
|
|
15
|
-
},
|
|
16
|
-
"stt": {
|
|
17
|
-
"$ref": "../components/recognizer",
|
|
18
|
-
"description": "Speech-to-text configuration for the pipeline."
|
|
19
|
-
},
|
|
20
|
-
"tts": {
|
|
21
|
-
"$ref": "../components/synthesizer",
|
|
22
|
-
"description": "Text-to-speech configuration for the pipeline."
|
|
23
|
-
},
|
|
24
|
-
"turnDetection": {
|
|
25
|
-
"oneOf": [
|
|
26
|
-
{
|
|
27
|
-
"type": "string",
|
|
28
|
-
"enum": ["stt", "krisp"],
|
|
29
|
-
"description": "Turn detection strategy shorthand. 'stt' uses the STT vendor's native signals (silence-based for most vendors; acoustic+semantic for deepgramflux, assemblyai, speechmatics). 'krisp' uses the Krisp acoustic end-of-turn model with default settings."
|
|
30
|
-
},
|
|
31
|
-
{
|
|
32
|
-
"type": "object",
|
|
33
|
-
"description": "Turn detection configuration with tunable parameters.",
|
|
34
|
-
"properties": {
|
|
35
|
-
"mode": {
|
|
36
|
-
"type": "string",
|
|
37
|
-
"enum": ["krisp"],
|
|
38
|
-
"description": "Turn detection mode. Currently only 'krisp' supports object-form tuning."
|
|
39
|
-
},
|
|
40
|
-
"threshold": {
|
|
41
|
-
"type": "number",
|
|
42
|
-
"minimum": 0,
|
|
43
|
-
"maximum": 1,
|
|
44
|
-
"description": "Krisp end-of-turn confidence threshold (0.0–1.0). Lower values trigger earlier turn transitions. Default: 0.5"
|
|
45
|
-
},
|
|
46
|
-
"model": {
|
|
47
|
-
"type": "string",
|
|
48
|
-
"description": "Optional Krisp model name override."
|
|
49
|
-
}
|
|
50
|
-
},
|
|
51
|
-
"required": ["mode"],
|
|
52
|
-
"additionalProperties": false
|
|
53
|
-
}
|
|
54
|
-
],
|
|
55
|
-
"default": "stt",
|
|
56
|
-
"description": "Turn detection strategy. Controls when the pipeline decides the user has finished speaking. STT vendors with native turn-taking (deepgramflux, assemblyai, speechmatics) always use their built-in detection regardless of this setting."
|
|
57
|
-
},
|
|
58
|
-
"bargeIn": {
|
|
59
|
-
"type": "object",
|
|
60
|
-
"description": "Controls whether and how the user can interrupt the assistant while it is speaking.",
|
|
61
|
-
"properties": {
|
|
62
|
-
"enable": {
|
|
63
|
-
"type": "boolean",
|
|
64
|
-
"description": "Allow the user to interrupt the assistant while it is speaking. Default: true.",
|
|
65
|
-
"default": true
|
|
66
|
-
},
|
|
67
|
-
"minSpeechDuration": {
|
|
68
|
-
"type": "number",
|
|
69
|
-
"minimum": 0,
|
|
70
|
-
"description": "Seconds of detected speech required before confirming an interruption. Prevents brief noises from cutting off the assistant. Default: 0.5",
|
|
71
|
-
"default": 0.5
|
|
72
|
-
},
|
|
73
|
-
"sticky": {
|
|
74
|
-
"type": "boolean",
|
|
75
|
-
"description": "If true, once the user interrupts the assistant does not resume speaking. Default: false.",
|
|
76
|
-
"default": false
|
|
77
|
-
}
|
|
78
|
-
},
|
|
79
|
-
"additionalProperties": false
|
|
80
|
-
},
|
|
81
|
-
"llm": {
|
|
82
|
-
"type": "object",
|
|
83
|
-
"description": "LLM configuration for the pipeline. See the 'llm' verb schema for details.",
|
|
84
|
-
"additionalProperties": true
|
|
85
|
-
},
|
|
86
|
-
"actionHook": {
|
|
87
|
-
"$ref": "../components/actionHook",
|
|
88
|
-
"description": "A webhook invoked when the pipeline ends."
|
|
89
|
-
},
|
|
90
|
-
"eventHook": {
|
|
91
|
-
"$ref": "../components/actionHook",
|
|
92
|
-
"description": "A webhook invoked for pipeline events. Receives event types: 'user_transcript' (user speech recognized), 'agent_response' (assistant reply), 'user_interruption' (barge-in detected), and 'turn_end' (end-of-turn summary with transcript, response, and latency metrics)."
|
|
93
|
-
},
|
|
94
|
-
"toolHook": {
|
|
95
|
-
"$ref": "../components/actionHook",
|
|
96
|
-
"description": "A webhook invoked when the LLM requests a tool/function call. The payload includes the tool name and arguments; the response provides the tool result."
|
|
97
|
-
},
|
|
98
|
-
"greeting": {
|
|
99
|
-
"type": "boolean",
|
|
100
|
-
"description": "Whether the LLM should generate an initial greeting before the user speaks. Default: true.",
|
|
101
|
-
"default": true
|
|
102
|
-
},
|
|
103
|
-
"earlyGeneration": {
|
|
104
|
-
"type": "boolean",
|
|
105
|
-
"description": "Enable speculative LLM prompting before end-of-turn is confirmed. When using Krisp turn detection, set this to true to speculatively prompt the LLM before Krisp confirms the turn has ended. If the transcript matches when turn ends, buffered tokens are released immediately — reducing response latency. Note: Deepgram Flux performs early generation automatically via its native EagerEndOfTurn signal regardless of this setting. Default: false.",
|
|
106
|
-
"default": false
|
|
107
|
-
},
|
|
108
|
-
"noiseIsolation": {
|
|
109
|
-
"oneOf": [
|
|
110
|
-
{
|
|
111
|
-
"type": "string",
|
|
112
|
-
"enum": ["krisp", "rnnoise"],
|
|
113
|
-
"description": "Shorthand — enable noise isolation with the specified vendor using default settings."
|
|
114
|
-
},
|
|
115
|
-
{
|
|
116
|
-
"type": "object",
|
|
117
|
-
"description": "Detailed noise isolation configuration.",
|
|
118
|
-
"properties": {
|
|
119
|
-
"mode": {
|
|
120
|
-
"type": "string",
|
|
121
|
-
"description": "Noise isolation vendor/mode (e.g. 'krisp')."
|
|
122
|
-
},
|
|
123
|
-
"level": {
|
|
124
|
-
"type": "number",
|
|
125
|
-
"minimum": 0,
|
|
126
|
-
"maximum": 100,
|
|
127
|
-
"description": "Suppression level 0–100. Default: 100."
|
|
128
|
-
},
|
|
129
|
-
"direction": {
|
|
130
|
-
"type": "string",
|
|
131
|
-
"enum": ["read", "write"],
|
|
132
|
-
"description": "Audio direction to apply noise isolation. 'read' filters caller audio, 'write' filters outbound audio. Default: 'read'."
|
|
133
|
-
},
|
|
134
|
-
"model": {
|
|
135
|
-
"type": "string",
|
|
136
|
-
"description": "Optional model name override."
|
|
137
|
-
}
|
|
138
|
-
},
|
|
139
|
-
"required": ["mode"],
|
|
140
|
-
"additionalProperties": false
|
|
141
|
-
}
|
|
142
|
-
],
|
|
143
|
-
"description": "Enable server-side noise isolation to reduce background noise on call audio. Defaults to filtering inbound (caller) audio; set direction to 'write' for outbound. Useful for improving STT accuracy in noisy environments."
|
|
144
|
-
},
|
|
145
|
-
"mcpServers": {
|
|
146
|
-
"type": "array",
|
|
147
|
-
"items": {
|
|
148
|
-
"type": "object",
|
|
149
|
-
"properties": {
|
|
150
|
-
"url": {
|
|
151
|
-
"type": "string",
|
|
152
|
-
"format": "uri",
|
|
153
|
-
"description": "The URL of the MCP server."
|
|
154
|
-
},
|
|
155
|
-
"auth": {
|
|
156
|
-
"type": "object",
|
|
157
|
-
"description": "Authentication for the MCP server.",
|
|
158
|
-
"additionalProperties": true
|
|
159
|
-
},
|
|
160
|
-
"roots": {
|
|
161
|
-
"type": "array",
|
|
162
|
-
"items": { "type": "object" },
|
|
163
|
-
"description": "MCP root definitions."
|
|
164
|
-
}
|
|
165
|
-
},
|
|
166
|
-
"required": ["url"]
|
|
167
|
-
},
|
|
168
|
-
"description": "External MCP servers that provide tools to the LLM. The pipeline connects at startup via SSE, discovers available tools, and makes them callable by the LLM."
|
|
169
|
-
}
|
|
170
|
-
},
|
|
171
|
-
"required": [
|
|
172
|
-
"llm"
|
|
173
|
-
],
|
|
174
|
-
"examples": [
|
|
175
|
-
{
|
|
176
|
-
"verb": "pipeline",
|
|
177
|
-
"stt": {
|
|
178
|
-
"vendor": "deepgram",
|
|
179
|
-
"language": "en-US"
|
|
180
|
-
},
|
|
181
|
-
"tts": {
|
|
182
|
-
"vendor": "cartesia",
|
|
183
|
-
"voice": "sonic-english"
|
|
184
|
-
},
|
|
185
|
-
"llm": {
|
|
186
|
-
"vendor": "openai",
|
|
187
|
-
"model": "gpt-4o",
|
|
188
|
-
"llmOptions": {
|
|
189
|
-
"messages": [
|
|
190
|
-
{
|
|
191
|
-
"role": "system",
|
|
192
|
-
"content": "You are a helpful voice assistant."
|
|
193
|
-
}
|
|
194
|
-
]
|
|
195
|
-
}
|
|
196
|
-
},
|
|
197
|
-
"turnDetection": "stt",
|
|
198
|
-
"actionHook": "/pipeline-complete"
|
|
199
|
-
},
|
|
200
|
-
{
|
|
201
|
-
"verb": "pipeline",
|
|
202
|
-
"stt": {
|
|
203
|
-
"vendor": "deepgram",
|
|
204
|
-
"language": "en-US"
|
|
205
|
-
},
|
|
206
|
-
"tts": {
|
|
207
|
-
"vendor": "cartesia",
|
|
208
|
-
"voice": "sonic-english"
|
|
209
|
-
},
|
|
210
|
-
"llm": {
|
|
211
|
-
"vendor": "anthropic",
|
|
212
|
-
"model": "claude-opus-4-6",
|
|
213
|
-
"llmOptions": {
|
|
214
|
-
"messages": [
|
|
215
|
-
{
|
|
216
|
-
"role": "user",
|
|
217
|
-
"content": "You are a helpful voice assistant."
|
|
218
|
-
}
|
|
219
|
-
]
|
|
220
|
-
}
|
|
221
|
-
},
|
|
222
|
-
"turnDetection": {
|
|
223
|
-
"mode": "krisp",
|
|
224
|
-
"threshold": 0.3
|
|
225
|
-
},
|
|
226
|
-
"bargeIn": {
|
|
227
|
-
"enable": true,
|
|
228
|
-
"minSpeechDuration": 0.3,
|
|
229
|
-
"sticky": false
|
|
230
|
-
},
|
|
231
|
-
"actionHook": "/pipeline-complete"
|
|
232
|
-
}
|
|
233
|
-
]
|
|
234
|
-
}
|