96Astro API

Documentation

Audio reports

Turn any report into a spoken MP3 in 26 languages, with your choice of voice, tone and length.

You can also create one, listen to it and download it in your dashboard, under PDF and audio jobs.

POST /v1/jobs/audio turns any report capability into a narrated MP3. It runs as a job, exactly like a PDF: submit it, poll it, then download the file.

curl https://api.96astroapi.com/v1/jobs/audio   -H "Authorization: Bearer $ASTRO_API_KEY"   -H "Content-Type: application/json"   -d '{
    "capability": "report-career",
    "input": { "datetime": "1990-10-18T14:30:00", "timezone": "Asia/Kolkata", "latitude": 28.6139, "longitude": 77.209 },
    "language": "hi",
    "minutes": 5,
    "tone": "warm",
    "voice": "meera",
    "listener_name": "Asha"
  }'
FieldRequiredMeaning
capabilityYesAny report capability, such as report-career, report-marriage, classical-kundli-report or matching-report
inputYesThat capability's request body
languageNoLanguage of the narration, default en. Any of the 26
minutesNoAbout how long it should be, 1 to 30. Default 5
toneNowarm (default), calm, formal or friendly
voiceNomeera (default), arjun, tara, dev, isha or kabir
listener_nameNoUsed to address the listener

What it costs: the report's own credits, plus 400 credits for each minute you ask for. If the job fails, nothing is charged.

How it is made: the report is calculated, a model writes a spoken script using only what the report says, and a speech model reads it out. The narration is AI-written and AI-spoken.

Sandbox keys get a one-minute narration, free, within the sandbox's daily job limit.

The file stays available for the same time as a PDF. Poll GET /v1/jobs/{id} for a fresh download_url.

What a finished job tells you

A finished audio job has an audio object: seconds (the real length of the file), voice, tone, language and minutes_requested.

Hear the voices

The Audio Report API page has a short sample of each of the six voices.

On this page