@gotcos/glasses-server 6.65.0 → 6.66.0

package/.env.example

@@ -340,3 +340,8 @@ BIND_HOST=0.0.0.0
 # additionally requires a client that acknowledges displaying continue_note.
 # Keep this in ~/.cos-glasses/.env; Control does not yet persist it on update.
 COS_CONTINUE_FULL_PERMISSIONS=1
+
+# Optional absolute executable paths for COS-managed whisper.cpp.
+# Unset uses Homebrew then PATH. An invalid explicit path fails closed.
+# COS_WHISPER_CLI_BIN=/absolute/path/to/whisper-cli
+# COS_WHISPER_SERVER_BIN=/absolute/path/to/whisper-server

package/CHANGELOG.md

@@ -1,3 +1,12 @@
+## 6.66.0 — 2026-10-08 — Managed voice setup and Haiku polish
+
+- Advertise Haiku, Sonnet and GPT-5.6 Luna Fast. Default omitted-model polish to Haiku for lower cost, while honoring explicit model choices and retaining the transcript on failure.
+- Glasses app 6.10.616 explicitly sends the chosen model and already works with server 6.65.0.
+- Setup, live preview, committed transcription and HQ share explicit `COS_WHISPER_CLI_BIN` / `COS_WHISPER_SERVER_BIN` resolution. A broken explicit path fails closed; existing Homebrew/PATH detection remains the fallback.
+- COS Control can prepare Turbo first for a local benchmark, then the chosen tier, without overwriting saved voice settings. Both preparation flags require `--setup-transcription --prepare-only`.
+- Pairs with COS Control 0.5.273 and the signed/notarized whisper.cpp 1.9.1 runtime.
+- Claude health probes resolve the installed absolute binary path, including managed-service environments without the interactive shell PATH.
+
 ## 6.65.0 — 2026-10-07 — Work search by meaning, and task dates
 
 - `POST /api/work/search` `{ query, domain?, scope?, ids? }` finds the Work card the user means, not only the words they typed. Pairs with COS Control 0.5.259, which shows its own word matches at once and asks this after a pause.

package/README.md

@@ -646,18 +646,24 @@ and counts only, never a speaker name, path or error text; those stay in
 
 ## Final dictation polish (6.61.5)
 
-With glasses app 6.10.573, **Settings → Voice → Polish model** offers **Sonnet**
-(the default, using your Claude CLI sign-in) and **GPT-5.6 Luna Fast** (using your
-Cursor CLI sign-in). Luna uses `gpt-5.6-luna-none-fast`; it needs server 6.61.5.
-Both phone text and recovered audio drafts honor the choice. The server advertises
-these options in `capabilities.dictationCleanup.models` on health and models.
+With glasses app 6.10.616, **Settings → Voice → Polish model** offers **Haiku 5.5**
+(the lower-cost default), **Sonnet 5.5** (higher-quality polish), and **GPT-5.6 Luna Fast**.
+Haiku and Sonnet use your Claude CLI sign-in; Luna uses your Cursor CLI sign-in.
+Luna uses `gpt-5.6-luna-none-fast` and needs server 6.61.5. Existing explicit choices
+are preserved. Unset or unknown app preferences use Haiku. Both phone text and
+recovered audio drafts honor the choice. Existing server 6.65.0 already accepts
+explicit Haiku requests, so the new app does not require a server update.
+
+The next server release also defaults omitted-model requests to Haiku and advertises
+all three options in `capabilities.dictationCleanup.models`. Until that release,
+6.65.0 still defaults requests without a model to Sonnet; 6.10.616 sends its choice
+explicitly. This is a polish-model change, not a change to the speech recognition engine.
 
 Cleanup uses subscription CLIs only, with no API-key fallback or automatic model
 substitution. A failed or unavailable cleanup keeps the glossary-corrected
 transcript. Turn **Final polish (AI)** off to skip the model call. Claude tools
 and hooks are disabled; Cursor runs in read-only ask mode with a temporary profile
 and empty workspace. Live preview and canonical meeting transcription are unchanged.
-Older clients may still explicitly request Haiku.
 
 ## HQ dictation
 
@@ -917,3 +923,23 @@ This is a local owner instance, separate from app.gotcos.com tenants and hosted
 ## Provider parity in 6.62.0
 
 See [provider capabilities and limits](docs/provider-parity.md) for Continue permissions, client compatibility, hook trust, observation limits and rollback. Glasses 6.10.600 shows the server's Continue note. Older clients retain Cursor Ask behavior; forks remain read-only.
+
+### Managed Whisper runtime (6.66.0 candidate)
+
+COS Control 0.5.272 prepares a signed, notarized whisper.cpp 1.9.1 runtime during
+voice setup on Apple silicon. It supplies absolute `COS_WHISPER_CLI_BIN` and
+`COS_WHISPER_SERVER_BIN` paths to the server; Homebrew is not needed for that flow.
+Direct server installs still fall back to Homebrew and then PATH when overrides
+are unset. A configured but missing/non-executable override is an error, never a
+silent change of runtime.
+
+`--setup-transcription --prepare-only --voice-benchmark-prepare` prepares Turbo
+without changing saved tier keys. Control measures the device and follows with
+`--setup-transcription --prepare-only --transcription-tier balanced|max
+--preserve-voice-settings`. The latter prepares all required models, including
+Large-v3 for saved-meeting HQ in both tiers. Applying a tier remains Control's
+transactional action. Downloading or cancelling preparation does not change it.
+Runtime/model setup is local; the existing provider sign-in requirements remain.
+
+Candidate only: publish the versioned runtime asset and server package before
+promoting the matching Control appcast entry.

package/bin/cli.cjs

@@ -25,7 +25,12 @@ const { homedir } = require('os')
 const PKG_ROOT = resolve(__dirname, '..')
 const CONFIG_DIR = join(homedir(), '.cos-glasses')
 const PREPARE_ONLY = process.argv.includes('--prepare-only')
+const VOICE_BENCHMARK_PREPARE = process.argv.includes('--voice-benchmark-prepare')
 const SETUP_TRANSCRIPTION = process.argv.includes('--setup-transcription')
+if ((VOICE_BENCHMARK_PREPARE || process.argv.includes('--preserve-voice-settings')) && !(SETUP_TRANSCRIPTION && PREPARE_ONLY)) {
+  console.error('Voice preparation flags require --setup-transcription --prepare-only.');
+  process.exit(64)
+}
 const SETUP_SPEAKER_MODEL = process.argv.includes('--setup-speaker-model')
 const HOOKS_ACTION = process.argv.includes('--hooks') ? process.argv[process.argv.indexOf('--hooks') + 1] : undefined
 function optionValue(name) {
@@ -540,12 +545,15 @@ function upsertEnvValue(file, key, value) {
   chmodSync(file, 0o600)
 }
 
-if (SETUP_TRANSCRIPTION) {
+if (SETUP_TRANSCRIPTION && !VOICE_BENCHMARK_PREPARE) {
+  const preserveVoiceSettings = process.argv.includes('--preserve-voice-settings')
   const previewModel = TRANSCRIPTION_TIER === 'max' ? 'turbo' : 'small.en'
   const commitModel = TRANSCRIPTION_TIER === 'max' ? 'large-v3' : 'turbo'
-  upsertEnvValue(ENV_FILE, 'COS_WHISPER_TRANSCRIPTION_TIER', TRANSCRIPTION_TIER)
-  upsertEnvValue(ENV_FILE, 'COS_WHISPER_PREVIEW_MODEL', previewModel)
-  upsertEnvValue(ENV_FILE, 'COS_WHISPER_COMMIT_MODEL', commitModel)
+  if (!preserveVoiceSettings) {
+    upsertEnvValue(ENV_FILE, 'COS_WHISPER_TRANSCRIPTION_TIER', TRANSCRIPTION_TIER)
+    upsertEnvValue(ENV_FILE, 'COS_WHISPER_PREVIEW_MODEL', previewModel)
+    upsertEnvValue(ENV_FILE, 'COS_WHISPER_COMMIT_MODEL', commitModel)
+  }
   process.env.COS_WHISPER_TRANSCRIPTION_TIER = TRANSCRIPTION_TIER
   process.env.COS_WHISPER_PREVIEW_MODEL = previewModel
   process.env.COS_WHISPER_COMMIT_MODEL = commitModel
@@ -573,8 +581,7 @@ if (!process.env.COS_PROFILE_PATH) process.env.COS_PROFILE_PATH = PROFILE_FILE
 
 // Step 5: local Whisper detection + model download. Voice stays local-only by
 // default; cloud fallback requires an explicit flag plus a configured key.
-const WHISPER_KNOWN_PATHS = ['/opt/homebrew/bin/whisper-cli', '/usr/local/bin/whisper-cli']
-const WHISPER_SERVER_KNOWN_PATHS = ['/opt/homebrew/bin/whisper-server', '/usr/local/bin/whisper-server']
+const { resolveWhisperBinary } = require('./whisper-runtime.cjs')
 const WHISPER_MODEL_DIR = join(homedir(), '.local/share/whisper-models')
 const WHISPER_MODEL_PATH = join(WHISPER_MODEL_DIR, 'ggml-large-v3-turbo.bin')
 const WHISPER_MODEL_PARTIAL = WHISPER_MODEL_PATH + '.partial'
@@ -591,24 +598,8 @@ const WHISPER_LARGE_MODEL_MIN_BYTES = 2_800_000_000
 const WHISPER_MODEL_EXPECTED_BYTES = 1_624_555_275
 const WHISPER_SMALL_MODEL_EXPECTED_BYTES = 487_614_201
 const WHISPER_LARGE_MODEL_EXPECTED_BYTES = 3_095_033_483
-function findWhisperCli() {
-  for (const p of WHISPER_KNOWN_PATHS) { if (existsSync(p)) return p }
-  try {
-    const found = execSync('command -v whisper-cli 2>/dev/null', { shell: '/bin/sh', stdio: 'pipe', timeout: 2000 }).toString().trim()
-    return found || null
-  } catch {
-    return null
-  }
-}
-function findWhisperServer() {
-  for (const p of WHISPER_SERVER_KNOWN_PATHS) { if (existsSync(p)) return p }
-  try {
-    const found = execSync('command -v whisper-server 2>/dev/null', { shell: '/bin/sh', stdio: 'pipe', timeout: 2000 }).toString().trim()
-    return found || null
-  } catch {
-    return null
-  }
-}
+function findWhisperCli() { return resolveWhisperBinary('whisper-cli') }
+function findWhisperServer() { return resolveWhisperBinary('whisper-server') }
 function isValidWhisperModel(p, minBytes = WHISPER_MODEL_MIN_BYTES) {
   if (!existsSync(p)) return false
   try { return statSync(p).size >= minBytes } catch { return false }
@@ -652,10 +643,10 @@ if (SETUP_TRANSCRIPTION && process.env.SKIP_WHISPER_DOWNLOAD !== '1') {
   mkdirSync(WHISPER_MODEL_DIR, { recursive: true })
   const targets = [
     [WHISPER_MODEL_PATH, WHISPER_MODEL_PARTIAL, WHISPER_MODEL_MIN_BYTES, WHISPER_MODEL_EXPECTED_BYTES],
-    ...(TRANSCRIPTION_TIER === 'balanced'
+    ...(!VOICE_BENCHMARK_PREPARE && TRANSCRIPTION_TIER === 'balanced'
       ? [[WHISPER_SMALL_MODEL_PATH, WHISPER_SMALL_MODEL_PARTIAL, WHISPER_SMALL_MODEL_MIN_BYTES, WHISPER_SMALL_MODEL_EXPECTED_BYTES]]
       : []),
-    [WHISPER_LARGE_MODEL_PATH, WHISPER_LARGE_MODEL_PARTIAL, WHISPER_LARGE_MODEL_MIN_BYTES, WHISPER_LARGE_MODEL_EXPECTED_BYTES],
+    ...(!VOICE_BENCHMARK_PREPARE ? [[WHISPER_LARGE_MODEL_PATH, WHISPER_LARGE_MODEL_PARTIAL, WHISPER_LARGE_MODEL_MIN_BYTES, WHISPER_LARGE_MODEL_EXPECTED_BYTES]] : []),
   ]
   const missingBytes = targets.reduce((sum, [path, , minimum, expected]) =>
     sum + (isValidWhisperModel(path, minimum) ? 0 : expected), 0)
@@ -716,7 +707,7 @@ if (process.env.COS_OPENAI_WHISPER_FALLBACK === '1') {
 
 const previewChoice = (process.env.COS_WHISPER_PREVIEW_MODEL || process.env.COS_WHISPER_REALTIME_MODEL || 'auto').trim().toLowerCase()
 const wantsSmallPreview = ['small', 'small.en', 'ggml-small.en.bin'].includes(previewChoice)
-if (whisperCliPath && wantsSmallPreview) {
+if (whisperCliPath && wantsSmallPreview && !VOICE_BENCHMARK_PREPARE) {
   if (isValidWhisperModel(WHISPER_SMALL_MODEL_PATH, WHISPER_SMALL_MODEL_MIN_BYTES)) {
     console.log(green('  ✓') + ' Small.en preview model ready ' + dim('— Turbo remains authoritative'))
   } else {
@@ -745,7 +736,7 @@ if (whisperCliPath && wantsSmallPreview) {
   }
 }
 
-if (whisperCliPath && SETUP_TRANSCRIPTION) {
+if (whisperCliPath && SETUP_TRANSCRIPTION && !VOICE_BENCHMARK_PREPARE) {
   if (isValidWhisperModel(WHISPER_LARGE_MODEL_PATH, WHISPER_LARGE_MODEL_MIN_BYTES)) {
     const largeRole = TRANSCRIPTION_TIER === 'max'
       ? 'live preview + commit and saved-meeting polish'
@@ -789,7 +780,7 @@ if (whisperCliPath && SETUP_TRANSCRIPTION) {
 // -- sixty times larger -- downloaded here with SKIP_WHISPER_DOWNLOAD as the
 // escape hatch. Same shape, same opt-out, so nobody on a metered or restricted
 // network is forced into it.
-if (!SETUP_SPEAKER_MODEL && process.env.SKIP_SPEAKER_MODEL_DOWNLOAD !== '1') {
+if (!VOICE_BENCHMARK_PREPARE && !SETUP_SPEAKER_MODEL && process.env.SKIP_SPEAKER_MODEL_DOWNLOAD !== '1') {
   const spkDest = join(CONFIG_DIR, 'models', SPEAKER_MODEL.filename)
   if (existsSync(spkDest) && sha256File(spkDest) === SPEAKER_MODEL.sha256) {
     console.log(green('  ✓') + ' Voiceprint model ready ' + dim('— named speakers available'))

package/bin/whisper-runtime.cjs

@@ -0,0 +1,17 @@
+// Shared by setup and both live Whisper workers. Explicit overrides fail closed.
+const { accessSync, constants, statSync } = require('node:fs')
+const { isAbsolute, join } = require('node:path')
+function executable(path) {
+  try { accessSync(path, constants.X_OK); return statSync(path).isFile() } catch { return false }
+}
+function resolveWhisperBinary(name, env = process.env, available = executable) {
+  if (!['whisper-cli', 'whisper-server'].includes(name)) throw new Error('Unknown Whisper binary')
+  const key = name === 'whisper-cli' ? 'COS_WHISPER_CLI_BIN' : 'COS_WHISPER_SERVER_BIN'
+  if (env[key] !== undefined && env[key] !== '') {
+    const path = env[key].trim()
+    return isAbsolute(path) && available(path) ? path : null
+  }
+  const candidates = ['/opt/homebrew/bin', '/usr/local/bin', ...(env.PATH || '').split(':')]
+  return candidates.filter(isAbsolute).map(p => join(p, name)).find(available) || null
+}
+module.exports = { resolveWhisperBinary }

package/bin/whisper-runtime.d.cts

@@ -0,0 +1 @@
+export function resolveWhisperBinary(name: 'whisper-cli' | 'whisper-server', env?: NodeJS.ProcessEnv, available?: (path: string) => boolean): string | null

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gotcos/glasses-server",
-  "version": "6.65.0",
+  "version": "6.66.0",
   "description": "COS Glasses: self-hosted AI heads-up-display server for Even G2 smart glasses, powered by Claude Code, Codex, Cursor Agent CLI, or local Ollama",
   "type": "module",
   "bin": {
@@ -77,7 +77,9 @@
     "docs/work-task-runtime.md",
     "!server/scripts/build-work-task-runtime.py",
     "bin/hooks/cos-cursor-observer.cjs",
-    "docs/provider-parity.md"
+    "docs/provider-parity.md",
+    "bin/whisper-runtime.cjs",
+    "bin/whisper-runtime.d.cts"
   ],
   "author": "COS Contributors",
   "license": "MIT",

package/server/lib/dictation-clean-models.ts

@@ -1,6 +1,6 @@
-/** Public choices. Haiku remains accepted only for older clients. */
-export const DICTATION_CLEAN_MODELS = ['sonnet', 'luna-5.6-fast'] as const
-export type DictationCleanModel = typeof DICTATION_CLEAN_MODELS[number] | 'haiku'
+/** Public polish choices. Haiku is the cost-first default. */
+export const DICTATION_CLEAN_MODELS = ['haiku', 'sonnet', 'luna-5.6-fast'] as const
+export type DictationCleanModel = typeof DICTATION_CLEAN_MODELS[number]
 export const LUNA_DICTATION_CLI_MODEL = 'gpt-5.6-luna-none-fast'
 export function isDictationCleanModel(value: unknown): value is DictationCleanModel {
   return value === 'sonnet' || value === 'luna-5.6-fast' || value === 'haiku'

package/server/lib/dictation-clean.ts

@@ -32,7 +32,7 @@ export async function autoCleanDictation(
   opts: { model?: string; signal?: AbortSignal } = {},
 ): Promise<string> {
   if (opts.signal?.aborted) throw new Error('Auto-clean aborted')
-  const model = (opts.model || process.env.COS_DICTATION_AUTOCLEAN_MODEL || 'sonnet').toLowerCase()
+  const model = (opts.model || process.env.COS_DICTATION_AUTOCLEAN_MODEL || 'haiku').toLowerCase()
   if (!isDictationCleanModel(model)) throw new Error('Unsupported dictation cleanup model')
   const cursor = model === 'luna-5.6-fast'
   const binary = resolveProviderBinary(cursor ? 'cursor' : 'claude')

package/server/lib/health-static-probes.ts

@@ -123,9 +123,11 @@ async function probePython(): Promise<string> {
   }
 }
 
-async function probeClaude(): Promise<{ value: string; available: boolean }> {
+export async function probeClaude(): Promise<{ value: string; available: boolean }> {
   try {
-    const result = await execute('claude', ['--version'])
+    const resolved = resolveProviderBinary('claude')
+    if (!resolved.ok) return { value: `unresolved (${resolved.detail})`, available: false }
+    const result = await execute(resolved.path, ['--version'])
     return { value: firstNonemptyLine(result.stdout) ?? 'available', available: true }
   } catch {
     return { value: 'error', available: false }

package/server/lib/whisper-local.ts

@@ -1,3 +1,4 @@
+import { resolveWhisperBinary } from '../../bin/whisper-runtime.cjs'
 // Local Whisper transcription via whisper-server (persistent) or whisper-cli (fallback)
 // Eliminates 1-3s OpenAI API round-trip by running inference on M3 Ultra locally.
 //
@@ -130,17 +131,9 @@ interface WhisperJsonResponse {
   text?: unknown
 }
 
-// Resolve whisper.cpp binaries across Homebrew prefixes (Apple Silicon
-// /opt/homebrew, Intel /usr/local). Downstream code existsSync-guards these
-// before use, so a missing binary degrades to CLI/cloud rather than crashing.
-function resolveWhisperBin(name: string): string {
-  for (const prefix of ['/opt/homebrew/bin', '/usr/local/bin']) {
-    if (existsSync(`${prefix}/${name}`)) return `${prefix}/${name}`
-  }
-  return `/opt/homebrew/bin/${name}`
-}
-const WHISPER_CLI = resolveWhisperBin('whisper-cli')
-const WHISPER_SERVER = resolveWhisperBin('whisper-server')
+// Same resolution as setup and preview; a broken explicit override never falls through to Homebrew.
+const WHISPER_CLI = resolveWhisperBinary('whisper-cli') ?? '/__cos_whisper_unavailable__/whisper-cli'
+const WHISPER_SERVER = resolveWhisperBinary('whisper-server') ?? '/__cos_whisper_unavailable__/whisper-server'
 const MODEL_DIR = join(process.env.HOME ?? homedir(), '.local/share/whisper-models')
 export const WHISPER_TURBO_MODEL_PATH = join(MODEL_DIR, 'ggml-large-v3-turbo.bin')
 export const WHISPER_LARGE_V3_MODEL_PATH = join(MODEL_DIR, 'ggml-large-v3.bin')

package/server/lib/whisper-preview.ts

@@ -1,3 +1,4 @@
+import { resolveWhisperBinary } from '../../bin/whisper-runtime.cjs'
 // Adaptive provisional transcription:
 //   Balanced -> isolated Small.en cosmetic preview + Turbo live commit
 //   Max      -> isolated Turbo cosmetic preview + Large-v3 live commit
@@ -55,8 +56,7 @@ export const WHISPER_SMALL_EN_MODEL_PATH = join(MODEL_DIR, 'ggml-small.en.bin')
 const WHISPER_TURBO_MODEL_PATH = join(MODEL_DIR, 'ggml-large-v3-turbo.bin')
 const VAD_MODEL_PATH = join(MODEL_DIR, 'ggml-silero-v5.1.2.bin')
 const VAD_ENABLED = process.env.COS_WHISPER_VAD !== '0'
-const WHISPER_SERVER = ['/opt/homebrew/bin/whisper-server', '/usr/local/bin/whisper-server']
-  .find(existsSync) ?? '/opt/homebrew/bin/whisper-server'
+const WHISPER_SERVER = resolveWhisperBinary('whisper-server') ?? '/__cos_whisper_unavailable__/whisper-server'
 const PREVIEW_PORT = 8177
 const PREVIEW_URL = `http://127.0.0.1:${PREVIEW_PORT}`
 const PROCESS_PROBE_TIMEOUT_MS = 2_000

package/server/routes/prompt-drafts.ts

@@ -136,7 +136,7 @@ async function cleanOutboundDictation(text: string, opts: AutoCleanRequest & { s
     console.warn(`[prompt-draft] Fuzzy correction failed (non-fatal): ${fuzzyErr?.message ?? fuzzyErr}`)
   }
   if (!(opts.enabled ?? autoCleanDefaultEnabled())) return cleaned
-  const model = opts.model ?? 'sonnet'
+  const model = opts.model ?? 'haiku'
   const autoCleanBreaker = autoCleanBreakers[model]
   if (cleaned.length > AUTOCLEAN_MAX_CHARS || autoCleanBreaker.isOpen() || autoCleanCountToday() >= autoCleanDailyCap()) return cleaned
   const startedAt = Date.now()