Audio

1
import { OpenRouter } from '@openrouter/sdk';
2
import fs from "fs/promises";
3
4
const openRouter = new OpenRouter({
5
  apiKey: '{{API_KEY_REF}}',
6
});
7
8
async function encodeAudioToBase64(audioPath: string): Promise<string> {
9
  const audioBuffer = await fs.readFile(audioPath);
10
  return audioBuffer.toString("base64");
11
}
12
13
// Read and encode the audio file
14
const audioPath = "path/to/your/audio.wav";
15
const base64Audio = await encodeAudioToBase64(audioPath);
16
17
const result = await openRouter.chat.send({
18
  model: "{{MODEL}}",
19
  messages: [
20
    {
21
      role: "user",
22
      content: [
23
        {
24
          type: "text",
25
          text: "Please transcribe this audio file.",
26
        },
27
        {
28
          type: "input_audio",
29
          inputAudio: {
30
            data: base64Audio,
31
            format: "wav",
32
          },
33
        },
34
      ],
35
    },
36
  ],
37
  stream: false,
38
});
39
40
console.log(result);

1
import requests
2
import json
3
import base64
4
5
url = "https://openrouter.ai/api/v1/chat/completions"
6
headers = {
7
    "Authorization": f"Bearer {API_KEY_REF}",
8
    "Content-Type": "application/json"
9
}
10
11
payload = {
12
    "model": "{{MODEL}}",
13
    "messages": [
14
        {
15
            "role": "user",
16
            "content": "Say hello in a friendly tone."
17
        }
18
    ],
19
    "modalities": ["text", "audio"],
20
    "audio": {
21
        "voice": "alloy",
22
        "format": "wav"
23
    },
24
    "stream": True
25
}
26
27
# Audio output requires streaming — the response is delivered as SSE chunks
28
response = requests.post(url, headers=headers, json=payload, stream=True)
29
30
audio_data_chunks = []
31
transcript_chunks = []
32
33
for line in response.iter_lines():
34
    if not line:
35
        continue
36
    decoded = line.decode("utf-8")
37
    if not decoded.startswith("data: "):
38
        continue
39
    data = decoded[len("data: "):]
40
    if data.strip() == "[DONE]":
41
        break
42
    chunk = json.loads(data)
43
    delta = chunk["choices"][0].get("delta", {})
44
    audio = delta.get("audio", {})
45
    if audio.get("data"):
46
        audio_data_chunks.append(audio["data"])
47
    if audio.get("transcript"):
48
        transcript_chunks.append(audio["transcript"])
49
50
transcript = "".join(transcript_chunks)
51
print(f"Transcript: {transcript}")
52
53
# Combine and decode the base64 audio chunks, then save
54
full_audio_b64 = "".join(audio_data_chunks)
55
audio_bytes = base64.b64decode(full_audio_b64)
56
with open("output.wav", "wb") as f:
57
    f.write(audio_bytes)