ElevenLabs Scribe v2 Realtime

Overview

Best for: Teams already using ElevenLabs TTS/agents who want the same vendor for live STT, including 8 kHz mu-law telephony.

At a glance

~150 ms vendor claim, conditions not stated. 90+ languages. Concurrency 9 is the Starter plan; Pro 30, Scale 45, Free 6. Free plan includes about 2.5 realtime hours. Diarization not listed for realtime. Keyterms +$0.08/hr. EU, India and Singapore residency hosts.

Features

• partial_transcript and committed_transcript
• commit_strategy: manual, vad, or turn_prediction
• VAD tuning (threshold, silence secs, min speech/silence ms)
• word timestamps (include_timestamps)
• language detection
• keyterms (up to 50, 20 chars each; paid)
• entity detection
• experimental transcript editing (paid)
• background audio filter
• enable_logging=false option
• regional and data-residency hosts

Pricing

Free tier: Free plan includes about 2.5 realtime hours.

Source: elevenlabs.io

Setup

1. Create an API key.
2. Server: connect to wss://api.elevenlabs.io/v1/speech-to-text/realtime?model_id=scribe_v2_realtime&audio_format=pcm_16000&commit_strategy=vad with header xi-api-key.
3. Browser: create a single-use token via the tokens endpoint on your server and pass ?token=.
4. Send input_audio_chunk JSON messages with base64 audio; read partial_transcript and committed_transcript.

Endpoint

wss://api.elevenlabs.io/v1/speech-to-text/realtime

Authentication

xi-api-key header, or single-use token in the token query parameter for client-side use.

import asyncio, base64, json, os, websockets  # pip install websockets>=14

URL = ("wss://api.elevenlabs.io/v1/speech-to-text/realtime"
       "?model_id=scribe_v2_realtime&audio_format=pcm_16000&commit_strategy=vad")

async def main():
    hdr = {"xi-api-key": os.environ["ELEVENLABS_API_KEY"]}
    async with websockets.connect(URL, additional_headers=hdr) as ws:
        async def send():
            with open("audio_16k_mono.raw", "rb") as f:
                while chunk := f.read(3200):  # 100 ms
                    await ws.send(json.dumps({
                        "message_type": "input_audio_chunk",
                        "audio_base_64": base64.b64encode(chunk).decode(),
                        "commit": False, "sample_rate": 16000}))
                    await asyncio.sleep(0.1)
        asyncio.create_task(send())
        async for msg in ws:
            d = json.loads(msg)
            t = d.get("message_type")
            if t == "partial_transcript":
                print("partial", d.get("text"))
            elif t == "committed_transcript":
                print("FINAL", d.get("text"))
            elif t and t.endswith("error"):
                print(d); break

asyncio.run(main())

Written from the current docs. Check the vendor's SDK version before you ship.

Warnings

Plus 12 warnings that apply to all speech-to-text, live APIs. See category warnings.

Limits

• Realtime concurrency limit by plan is in a separate chart on the models page (not captured).
• keepalive_interval_ms configurable 500-10000; errors include session_time_limit_exceeded, commit_throttled, queue_overflow, insufficient_audio_activity.

Models and products

Docs and sources

Docs

• Realtime API reference
• STT capability page
• API pricing

Sources used

• elevenlabs.io/pricing/api
• elevenlabs.io/docs/capabilities/speech-to-text
• elevenlabs.io/docs/api-reference/speech-to-text/v-1-speech-to-text-realtime