Add voice to your AI app with Gladia

Prerequisites

Step-by-Step

1. 1

   Get your API key

   Sign up at gladia.io and grab the key from the dashboard. Free tier covers most prototyping.

2. 2

   Transcribe a file

   Upload a file URL and poll for the result. Pre-recorded transcription returns timestamps, confidence, and full text.

   curl -X POST https://api.gladia.io/v2/transcription \
     -H 'x-gladia-key: $GLADIA_API_KEY' \
     -H 'Content-Type: application/json' \
     -d '{"audio_url": "https://example.com/recording.mp3"}'

3. 3

   Enable diarization

   Diarization labels each segment with a speaker ID. Indispensable for meeting recordings and podcasts.

   { "audio_url": "...", "diarization": true, "diarization_config": { "number_of_speakers": 2 } }

4. 4

   Stream from a microphone

   Open a websocket, push 16kHz PCM frames, receive partial + final transcripts. Latency under 300ms.

   const ws = new WebSocket('wss://api.gladia.io/audio/text/audio-transcription', { headers: { 'x-gladia-key': key } });
   ws.on('open', () => ws.send(JSON.stringify({ x_gladia_key: key, language: 'en', encoding: 'wav' })));
   ws.on('message', (d) => console.log(JSON.parse(d.toString())));

5. 5

   Use the gladia CLI

   For local files, the gladia CLI handles upload + polling in one command.

   gladia recording.mp4 -o transcript.md --diarization

6. 6

   Pipe into your app

   Take the transcript JSON, feed segments into a summarizer or RAG ingest pipeline, and you have searchable audio.

Common Pitfalls

What's Next

• ->Pair with ElevenLabs for full STT+TTS conversational agents.
• ->Add a summarization step to convert transcripts into action items.

More Add voice to your AI app