صدا: گفتار و متن

curl https://gisoo.pro/api/v1/audio/transcriptions \
  -H "Authorization: Bearer $GISOO_API_KEY" \
  -F file=@meeting.mp3 \
  -F model=gpt-4o-mini-transcribe \
  -F language=fa

from openai import OpenAI

client = OpenAI(base_url="https://gisoo.pro/api/v1", api_key="sk-gisoo-v1-...")

with open("meeting.mp3", "rb") as audio:
    result = client.audio.transcriptions.create(
        model="gpt-4o-mini-transcribe",
        file=audio,
        language="fa",
        prompt="نام‌ها: گیسو، نِت اَرز، لاراول",   # optional: spellings the model should know
    )
print(result.text)

import fs from "node:fs";

const result = await client.audio.transcriptions.create({
  model: "gpt-4o-mini-transcribe",
  file: fs.createReadStream("meeting.mp3"),
  language: "fa",
});
console.log(result.text);

<?php
$result = $client->audio()->transcribe([
    'model' => 'gpt-4o-mini-transcribe',
    'file' => fopen('meeting.mp3', 'r'),
    'language' => 'fa',
]);
echo $result->text;

# Subtitles with timestamps (whisper-1)
curl https://gisoo.pro/api/v1/audio/transcriptions \
  -H "Authorization: Bearer $GISOO_API_KEY" \
  -F file=@lecture.mp3 -F model=whisper-1 -F language=fa \
  -F response_format=srt > lecture.srt

with open("lecture.mp3", "rb") as audio:
    result = client.audio.transcriptions.create(
        model="whisper-1",
        file=audio,
        response_format="verbose_json",
        timestamp_granularities=["word", "segment"],
    )
for seg in result.segments:
    print(f"{seg.start:7.2f}–{seg.end:7.2f}  {seg.text}")

with open("interview-fa.mp3", "rb") as audio:
    english = client.audio.translations.create(model="whisper-1", file=audio)
print(english.text)   # English text of a Persian recording

# Mono, 32 kbps MP3 — speech stays clear, the file shrinks a lot
ffmpeg -i input.m4a -ac 1 -b:a 32k -ar 16000 small.mp3

# Cut a long recording into 15-minute pieces
ffmpeg -i long.mp3 -f segment -segment_time 900 -c copy part-%02d.mp3

import base64

with open("voice-note.mp3", "rb") as f:
    b64 = base64.b64encode(f.read()).decode()

reply = client.chat.completions.create(
    model="gemini-3.6-flash",
    max_tokens=4000,
    messages=[{
        "role": "user",
        "content": [
            {"type": "text", "text": "Transcribe this Persian recording with [mm:ss] timestamps every few sentences."},
            {"type": "input_audio", "input_audio": {"data": b64, "format": "mp3"}},
        ],
    }],
)
print(reply.choices[0].message.content)

curl https://gisoo.pro/api/v1/audio/speech \
  -H "Authorization: Bearer $GISOO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini-tts",
    "voice": "nova",
    "input": "Your order has shipped and will arrive within two days.",
    "response_format": "mp3"
  }' --output notice.mp3

speech = client.audio.speech.create(
    model="gpt-4o-mini-tts",
    voice="nova",
    input="Your order has shipped and will arrive within two days.",
    instructions="Warm and calm, like a helpful shop assistant.",   # gpt-4o-mini-tts only
)
speech.write_to_file("notice.mp3")

import fs from "node:fs";

const speech = await client.audio.speech.create({
  model: "gpt-4o-mini-tts",
  voice: "nova",
  input: "Your order has shipped and will arrive within two days.",
});
fs.writeFileSync("notice.mp3", Buffer.from(await speech.arrayBuffer()));