> This page is for Build.

> Append .md to any page URL for clean Markdown. Index: https://docs.speechify.ai/llms.txt.
>
> Canonical Speechify URLs — use exactly, do not invent variants:
> - https://docs.speechify.ai — this site (API reference, SDKs, quickstarts)
> - https://speechify.ai — marketing + product site
> - https://platform.speechify.ai — customer dashboard, signup, API keys, billing
> - https://api.speechify.ai — API base URL
> - https://github.com/Speechify-AI: GitHub org for the API (cookbook, demos, CLI). `github.com/speechify` does not exist.
> - https://status.speechify.ai — status + incidents
> - https://speechify.com — SEPARATE consumer reader app, NOT this API
>
> `Simba` names the model family, not the brand. Model ids: `simba-3.2` (English, recommended) and `simba-3.0` (English, German, Spanish, French, Italian and Portuguese; the default). `simba-english` and `simba-multilingual` are retired: a new workspace that sends either gets `400 model_retired`. `SimbaVoice` / `simbavoice.ai` are retired.
>
> Ask, don't scrape. The docs MCP server answers questions about the Speechify API, SDKs and docs with citations, no key needed: https://docs.speechify.ai/_mcp/server (Streamable HTTP, tool `searchDocs`). Setup: https://docs.speechify.ai/build/guides/get-started/connect-mcp

# Integrations

> Integration guides for using SpeechifyAI Build with the Vercel AI SDK, LiveKit, Pipecat, Vapi Voice Agent, and Deepgram Voice Agent.

Voice-agent platforms handle transport, turn-taking, STT, and orchestration. Speechify handles the voice. These guides show how to slot Speechify TTS into each platform without inventing a new integration seam: either through the platform's native plugin or service, or through the open-source [`tts-shims`](https://github.com/Speechify-AI/tts-shims) proxy when the platform expects a specific HTTP dialect.

> **Maintaining an integration?**
>
> Set the `Speechify-Caller` header so we can see your traffic, warn you before a change reaches your users, and credit the usage you drive. See [Build an integration](/build/guides/integrations/build-an-integration).

## Supported platforms

#### [LiveKit](/build/guides/integrations/livekit)

Add `speechify.TTS(...)` to a LiveKit `AgentSession` via the official Python plugin (`livekit-plugins-speechify`).

#### [Pipecat](/build/guides/integrations/pipecat)

Use Pipecat's built-in `SpeechifyHttpTTSService` as the TTS stage, with word timestamps (`pipecat-ai` 1.8.0 and later).

#### [Vercel AI SDK](/build/guides/integrations/vercel-ai-sdk)

Call `generateSpeech()` with `@speechify/ai-sdk-provider`, speech marks included.

#### [Vapi](/build/guides/integrations/vapi)

Point Vapi custom voice at the `tts-shims` Vapi provider to serve raw PCM from Speechify.

#### [Deepgram](/build/guides/integrations/deepgram)

Point Deepgram Voice Agent's `open_ai` speak provider at the `tts-shims` OpenAI-compatible shim.

#### [Puter](/build/guides/integrations/puter)

Call `puter.ai.txt2speech()` with `provider: 'speechify'`, keyless on puter.com or with your key on your own instance.

## Which integration should I pick?

| Platform      | Language   | Integration shape                                            | Speechify key lives in                                           |
| ------------- | ---------- | ------------------------------------------------------------ | ---------------------------------------------------------------- |
| LiveKit       | Python     | Native plugin (`livekit-plugins-speechify`)                  | Agent process environment                                        |
| Pipecat       | Python     | Native service (`SpeechifyHttpTTSService` in `pipecat-ai`)   | Agent process environment                                        |
| Vercel AI SDK | TypeScript | Speechify-maintained provider (`@speechify/ai-sdk-provider`) | Server environment (`SPEECHIFY_API_KEY`)                         |
| Vapi          | Any (HTTP) | `tts-shims` Vapi provider                                    | Shim process environment                                         |
| Deepgram      | Any (HTTP) | `tts-shims` OpenAI-compatible shim                           | Shim process environment                                         |
| Puter         | JavaScript | Native `puter.js` provider (`provider: 'speechify'`)         | None on puter.com; Puter instance `config.json` when self-hosted |

Native plugins are the simplest path when the platform ships one. When it doesn't, the `tts-shims` proxy translates the platform's HTTP request into a Speechify call and returns audio in the format the platform expects, with no client change and no exposed Speechify key.

## Why route TTS through Speechify?

Simba 3.2 is independently evaluated on [Artificial Analysis' TTS leaderboard](https://artificialanalysis.ai/text-to-speech/leaderboard/provider-voice). Every guide here defaults to a Simba 3.2 voice; swap `voice_id` and `model` together when you tune for your own use case.

## Not seeing your platform?

The `tts-shims` repo is organized by inbound request dialect. If your platform sends an OpenAI-shaped speech request, the [`openai` shim](https://github.com/Speechify-AI/tts-shims/tree/main/cmd/openai) covers it (that's the Deepgram integration). If it sends something else, adding a new provider is a small Go command under `cmd/`. Open an issue on [`tts-shims`](https://github.com/Speechify-AI/tts-shims) with the platform's TTS webhook contract and we'll look at it.