Text to speech API for real-time voice

1. 1

   Get an API key

   Create a free account and copy an API key from your dashboard. New accounts include 50k characters of speech, so every step below is free to try.

2. 2

   Pick a voice

   Every request names a voice by UUID. List what your account can use, or start with Karen, one of three built-in voices.

   List voices

   curl "https://api.bland.ai/v1/voices" \
     -H "Authorization: Bearer $BLAND_API_KEY"

3. 3

   Generate a WAV file

   Text and voice are all you need. The pause marker adds 0.4 seconds of silence, and the two controls shape the read.

   cURL

   curl -X POST "https://api.bland.ai/v2/tts" \
     -H "Authorization: Bearer $BLAND_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
       "text": "Thanks for calling. <|0.4|> How can I help you today?",
       "voice": "29158307-9893-4149-8a75-bc9ce313d64e",
       "controls": { "expressiveness": 0.8, "stability": 0.4 },
       "audio": { "container": "wav" }
     }' \
     --output hello.wav

4. 4

   Stream it

   Without a container, the response is 48 kHz PCM you can play as it lands, the rate the model renders natively. For LLM output, stream tokens into the WebSocket.

   HTTP streaming

   curl -N -X POST "https://api.bland.ai/v2/tts" \
     -H "Authorization: Bearer $BLAND_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
       "text": "Your appointment is confirmed for Tuesday at three.",
       "voice": "29158307-9893-4149-8a75-bc9ce313d64e",
       "audio": { "encoding": "pcm_s16le", "sample_rate": 48000 }
     }' > out.pcm

   WebSocket (Node)

   // npm install ws. llmTextStream is your model's token stream,
   // playback is your audio queue.
   import WebSocket from "ws";
   
   const ws = new WebSocket("wss://api.bland.ai/v2/tts/ws", {
     headers: { Authorization: "Bearer YOUR_API_KEY" },
   });
   const turn = crypto.randomUUID();
   
   ws.on("open", () => {
     ws.send(JSON.stringify({
       type: "init",
       voice: "29158307-9893-4149-8a75-bc9ce313d64e",
       audio: { encoding: "pcm_s16le", sample_rate: 48000 },
     }));
   });
   
   ws.on("message", async (data, isBinary) => {
     if (isBinary) return playback.enqueue(data); // raw PCM frames
   
     const msg = JSON.parse(data.toString());
     if (msg.type === "ready") {
       for await (const token of llmTextStream) {
         ws.send(JSON.stringify({ type: "speak", context_id: turn, text: token }));
       }
       ws.send(JSON.stringify({ type: "end_of_turn", context_id: turn }));
     }
     if (msg.type === "utterance_end") ws.send(JSON.stringify({ type: "close" }));
   });

The speech quickstart covers telephony formats, every field, and the full realtime example.