@pinecall/skills 0.1.8 → 0.1.10

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pinecall/skills",
-  "version": "0.1.8",
+  "version": "0.1.10",
   "description": "Agent Skills for the Pinecall SDK — installable into Claude Code, Antigravity, Cursor, Copilot and any agent that supports the open Skills format.",
   "type": "module",
   "license": "MIT",

package/skills/pinecall-reference/references/reference/managed-vs-byok.md

@@ -34,10 +34,12 @@ Every STT, TTS and LLM model on Pinecall is one of two kinds:
 
 | Service | BYOK-only providers |
 |---|---|
-| **STT** | `assemblyai` |
-| **TTS** | `rime` |
+| **STT** | `assemblyai`, `soniox` |
+| **TTS** | `rime`, `soniox` |
 | **LLM** | `xai` (grok), `groq`, `cerebras`, `deepseek`, `openrouter` |
 
+> `soniox` is one key for **both** STT and TTS (a Soniox key enables both).
+
 ## Check it from the API (authoritative, live)
 
 The rate table is the source of truth. Query it any time:
@@ -86,6 +88,26 @@ One key can cover multiple services where a provider shares it — e.g. an
 **ElevenLabs** key enables both ElevenLabs TTS and ElevenLabs Scribe STT; a
 **Cartesia** key enables Sonic TTS and Ink-Whisper STT.
 
+## Operators: flip a provider managed ↔ BYOK
+
+The managed/BYOK status of any provider is a single DB flag (`rate.managed`) and is
+the source of truth — flipping it takes effect **immediately** (next session / page
+load), no deploy or re-seed:
+
+```bash
+# Pinecall provides it (no token needed) — requires <PROVIDER>_API_KEY on the box:
+npx tsx scripts/set-managed.ts assemblyai true
+
+# Force bring-your-own-key (the user must add their token):
+npx tsx scripts/set-managed.ts assemblyai false
+
+# Per service (e.g. only Cartesia STT):
+npx tsx scripts/set-managed.ts cartesia false stt
+```
+
+Enabling managed only works if Pinecall actually holds that provider's key in the
+prod `.env` (`shipway env` to check); otherwise leave it BYOK.
+
 ## What's next
 
 - [STT Providers](/reference/stt-providers)

package/skills/pinecall-reference/references/reference/stt-providers.md

@@ -29,6 +29,7 @@ Pinecall supports multiple STT providers. Use the `provider/model` format or a f
 { stt: "cartesia/ink-whisper" }      // Cartesia Ink-Whisper
 { stt: "elevenlabs/scribe" }         // ElevenLabs Scribe v2 (realtime)
 { stt: "assemblyai/universal" }      // AssemblyAI Universal-3
+{ stt: "soniox/realtime" }           // Soniox real-time (BYOK)
 ```
 
 ## Managed vs bring-your-own-key (BYOK)
@@ -46,6 +47,7 @@ for the full list and the live `GET /api/rates/models` query.
 | `cartesia` (ink-whisper) | ✅ Yes | Same key as Cartesia TTS — Pinecall hosts it |
 | `elevenlabs` (scribe) | ✅ Yes | Same key as ElevenLabs TTS — Pinecall hosts it |
 | `assemblyai` (universal) | ❌ BYOK only | Add an AssemblyAI key |
+| `soniox` (realtime) | ❌ BYOK only | One Soniox key = STT **and** TTS |
 
 > **BYOK enforcement:** if you configure a BYOK-only STT provider and your org has
 > not saved a key for it, **agent registration is rejected** with
@@ -179,6 +181,17 @@ stt: {
 }
 ```
 
+## Soniox (BYOK)
+
+Real-time multilingual STT (60+ languages). One Soniox key serves **both** Soniox
+STT and TTS. Requires your own Soniox key.
+
+```typescript
+stt: "soniox/realtime"
+// or
+stt: { provider: "soniox", model: "stt-rt-v5", language: "en" }
+```
+
 ## Which to choose
 
 | Provider | Best for | Trade-off |
@@ -190,6 +203,7 @@ stt: {
 | `cartesia/ink-whisper` | Single-vendor with Cartesia TTS | Managed (shared key) |
 | `elevenlabs/scribe` | Single-vendor with ElevenLabs TTS | Managed (shared key) |
 | `assemblyai/universal` | Accuracy + diarization | BYOK only |
+| `soniox/realtime` | Multilingual (60+), single-vendor with Soniox TTS | BYOK only |
 
 For most agents, start with `deepgram/flux`. Use `deepgram/nova-3` for languages Flux doesn't cover (Arabic, Hindi, Thai, Chinese, Japanese, Korean, etc.).
 

package/skills/pinecall-reference/references/reference/tts-providers.md

@@ -29,9 +29,10 @@ for the full list and the live `GET /api/rates/models` query.
 | TTS provider | Managed (no key needed) | Notes |
 |---|---|---|
 | `elevenlabs` | ✅ Yes | Default, recommended |
-| `cartesia` (sonic) | ✅ Yes | |
+| `cartesia` (sonic-3.5) | ✅ Yes | |
 | `polly` (AWS) | ✅ Yes | |
 | `rime` | ❌ BYOK only | Add a Rime key under Provider Keys |
+| `soniox` | ❌ BYOK only | One Soniox key = TTS **and** STT |
 
 > **BYOK enforcement:** configuring `rime` without a saved Rime key rejects agent
 > registration with `PROVIDER_KEY_REQUIRED`. With your own key, that usage is billed
@@ -156,7 +157,7 @@ The model is part of the voice config, so it hot-reloads with it — `agent.upda
 voice: {
   provider: "cartesia",
   voice_id: "a0e99841-438c-4a64-b679-ae501e7d6091",
-  model: "sonic-3",
+  model: "sonic-3.5",   // latest; also "sonic-3" / "sonic-latest"
   speed: 1.0,
   volume: 1.0,
   emotion: null,
@@ -168,7 +169,7 @@ Shortcut: `"cartesia/yumiko"`
 
 **Tuning notes:**
 
-- `model: "sonic-3"` — fastest Cartesia model, designed for streaming
+- `model: "sonic-3.5"` — latest/fastest Cartesia model (sub-90ms, 42 languages), designed for streaming. `sonic-3` and `sonic-latest` also available.
 - `emotion` accepts named emotion presets (check Cartesia docs for the current list)
 
 ## AWS Polly
@@ -204,6 +205,22 @@ voice: {
 
 Shortcut: `"rime/cove"`
 
+## Soniox (BYOK)
+
+Real-time TTS in 60+ languages. One Soniox key serves **both** Soniox TTS and STT.
+Requires your own Soniox key.
+
+```typescript
+voice: {
+  provider: "soniox",
+  voice_id: "Adrian",   // Soniox voice name
+  model: "tts-rt-v1",
+  language: "en",
+}
+```
+
+Shortcut: `"soniox/Adrian"`
+
 ## Which to choose
 
 | Provider | Best for | Trade-off |
@@ -212,8 +229,9 @@ Shortcut: `"rime/cove"`
 | **Cartesia** | Real-time streaming, low latency | Smaller voice library |
 | **Polly** | Cheap IVR, simple flows | Less natural |
 | **Rime** | Ultra-natural expressive English | BYOK only; English-focused |
+| **Soniox** | Multilingual (60+), single-vendor with Soniox STT | BYOK only |
 
-For most agents, start with ElevenLabs (`eleven_flash_v2_5`) or Cartesia (`sonic-3`). Use Polly only for high-volume, low-engagement flows.
+For most agents, start with ElevenLabs (`eleven_flash_v2_5`) or Cartesia (`sonic-3.5`). Use Polly only for high-volume, low-engagement flows.
 
 ## Hot-reloading voices