> For the complete documentation index, see [llms.txt](https://docs.aimlapi.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aimlapi.com/api-references/realtime-models/xai/grok-voice-think-fast-2.0.md). # Grok Voice Think Fast 2.0 {% columns %} {% column width="66.66666666666666%" %} {% hint style="info" %} This documentation is valid for the following list of our models: * `x-ai/grok-voice-think-fast-2.0` * `grok-voice-think-fast-2.0` * `x-ai/grok-voice-latest` * `grok-voice-latest` {% endhint %} {% endcolumn %} {% column width="33.33333333333334%" %} Try in Playground {% endcolumn %} {% endcolumns %} xAI's voice agent model. It speaks the OpenAI Realtime event names with xAI's own session shape and is billed by session duration; `x-ai/grok-voice-latest` and `grok-voice-latest` always open a session on `x-ai/grok-voice-think-fast-2.0`. ## Setup your API Key If you don’t have an API key for the AI/ML API yet, feel free to use our [Quickstart guide](https://docs.aimlapi.com/quickstart/setting-up). ## API Schema The model is served over a WebSocket in the xAI realtime voice protocol. Client events, billing, limits and errors are described in [Realtime Sessions](/capabilities/realtime-sessions.md#xai-realtime-voice). | | | | -------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | Connection URL | `wss://api.aimlapi.com/v1/realtime?model=x-ai/grok-voice-think-fast-2.0` | | Authentication | Server: `Authorization: Bearer `. Browser: subprotocols `realtime` and `openai-insecure-api-key.` ([client secrets](/capabilities/realtime-sessions.md#browsers-client-secrets)) | Session options (`session.update`): | Option | Values | | ------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | `instructions` | Up to 16,384 characters | | `voice` | Built-in voices (case-insensitive): `eve`, `ara`, `leo`, `rex`, `sal`, `carina`, `zagan`, `helix`, `orion`, `luna`, `iris`, `altair`, `zenith`, `perseus`, `helios`, `lux`, `kepler`, `rigel`, `cosmo`, `celeste`, `ursa`, `sirius`, `lumen`, `castor`, `naksh`, `atlas` | | `turn_detection` | `{"type": "server_vad"}`, optionally with `idle_timeout_ms` (1,000 – 600,000) | | `reasoning.effort` | `high` or `none` | | `replace` | Up to 100 pronunciation replacements, for example `{"Acme Mobile": "Acme Mobull"}` | | `audio.input.format`, `audio.output.format` | `type`: `audio/pcm`, `audio/pcmu`, `audio/pcma`, `audio/opus`; `rate`: 8000, 11025, 16000, 22050, 24000, 32000, 44100 or 48000 | | `audio.input.transport`, `audio.output.transport` | `json` (base64 in events) or `binary` (raw audio in binary WebSocket frames) | | `audio.input.transcription` | `language_hint`, up to 100 `keyterms` | | `audio.output.speed` | 0.7 – 1.5 | | `tools` | Up to 32 function tools in xAI's shape: `{"type": "function", "function": {"name", "description", "parameters"}}` | ## Code Example {% tabs %} {% tab title="Python" %} {% code overflow="wrap" %} ```python # pip install websockets import asyncio import base64 import json import websockets # Insert your AI/ML API key instead of : API_KEY = "" URL = "wss://api.aimlapi.com/v1/realtime?model=x-ai/grok-voice-think-fast-2.0" # 100 ms of 24 kHz mono PCM16 silence; send real microphone audio instead. SILENCE = base64.b64encode(bytes(4800)).decode() async def keep_streaming(ws): while True: await ws.send(json.dumps({"type": "input_audio_buffer.append", "audio": SILENCE})) await asyncio.sleep(5) async def main(): async with websockets.connect( URL, additional_headers={"Authorization": f"Bearer {API_KEY}"} ) as ws: await ws.send(json.dumps({ "type": "session.update", "session": { "instructions": "You are a friendly assistant. Answer briefly.", "voice": "eve", "turn_detection": {"type": "server_vad"}, }, })) streamer = asyncio.create_task(keep_streaming(ws)) await ws.send(json.dumps({ "type": "conversation.item.create", "item": { "type": "message", "role": "user", "content": [{"type": "input_text", "text": "Say hello in five words."}], }, })) await ws.send(json.dumps({"type": "response.create"})) async for message in ws: event = json.loads(message) print(event["type"]) if event["type"] == "response.done": break if event["type"] == "error": print("Error:", event["error"]) break streamer.cancel() asyncio.run(main()) ``` {% endcode %} {% endtab %} {% tab title="JavaScript" %} {% code overflow="wrap" %} ```javascript // npm i ws import WebSocket from "ws"; // Insert your AI/ML API key instead of : const apiKey = ""; const ws = new WebSocket( "wss://api.aimlapi.com/v1/realtime?model=x-ai/grok-voice-think-fast-2.0", { headers: { Authorization: `Bearer ${apiKey}` } }, ); let timer; ws.on("open", () => { ws.send(JSON.stringify({ type: "session.update", session: { instructions: "You are a friendly assistant. Answer briefly.", voice: "eve", turn_detection: { type: "server_vad" }, }, })); // 100 ms of 24 kHz mono PCM16 silence every 5 s; send real microphone audio instead. const silence = Buffer.alloc(4800).toString("base64"); timer = setInterval( () => ws.send(JSON.stringify({ type: "input_audio_buffer.append", audio: silence })), 5_000, ); ws.send(JSON.stringify({ type: "conversation.item.create", item: { type: "message", role: "user", content: [{ type: "input_text", text: "Say hello in five words." }], }, })); ws.send(JSON.stringify({ type: "response.create" })); }); ws.on("message", (data) => { const event = JSON.parse(data.toString()); console.log(event.type); if (event.type === "response.done") { clearInterval(timer); ws.close(); } else if (event.type === "error") { console.error("Error:", event.error); } }); ``` {% endcode %} {% endtab %} {% endtabs %}
Response The example sends one text message (billed as one text item), appends a little audio every few seconds where a real client would stream the microphone, and prints the type of every server event until `response.done`. Values in angle brackets are placeholders. {% code overflow="wrap" %} ``` ... response.done ``` {% endcode %}
--- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation by asking a question. Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters: ``` GET https://docs.aimlapi.com/api-references/realtime-models/xai/grok-voice-think-fast-2.0.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.