Text-to-speech API · 11 languages · 20 voices
Early accessWords, out loud.
louds.ai speaks English and 10 Indian languages. Stream speech to a voice agent as it thinks, render a clip in one request, or narrate a whole book in one job — through one API, priced per character.
In private build
CH 01 — Languages
English and 10 Indian languages.
- EnglishEnglishen
- বাংলাBengalibn
- ગુજરાતીGujaratigu
- हिन्दीHindihi
- ಕನ್ನಡKannadakn
- മലയാളംMalayalamml
- मराठीMarathimr
- ଓଡ଼ିଆOdiaor
- ਪੰਜਾਬੀPunjabipa
- தமிழ்Tamilta
- తెలుగుTelugute
Every catalog voice is labelled with the languages it speaks well. Cloned voices speak all 12, Assamese included.
CH 02 — Endpoints
Three ways to speak.
- 01
Live
For voice agents, assistants, games
Audio starts while the rest is still being written. Stream over HTTP, or open a WebSocket and feed text in as your model produces it.
2,000 characters per request
- 02
Clip
For notifications, prompts, short reads
One request, one finished file — with sentence timings and captions if you want them.
5,000 characters per request
- 03
Long-form
For articles, courses, books
We split, render, stitch and level the audio, then ping your webhook. Chapters, SRT and VTT included.
1,000,000 characters per job
CH 03 — Your voice
Bring your own voice.
- 1Record 3 to 10 seconds of clear speech.
- 2Read one consent line, so we know it's you.
- 3Your voice speaks all 12 languages.
Private to your organization. Every file from a cloned voice is watermarked. Cloning anyone without their consent is not allowed.
CH 04 — Developers
Show the request. Get the audio.
- snake_case JSON, integer milliseconds, speed as a plain multiplier.
- Sentence timings and captions — SRT and VTT — on any request.
- OpenAI-compatible endpoint, so existing clients switch with a base URL.
- Phone-ready audio: 8 kHz µ-law and 16 kHz PCM.
curl https://api.louds.ai/v1/tts \
-H "Authorization: Bearer $LOUDS_API_KEY" \
-H "Content-Type: application/json" \
-d '{"text":"Your order shipped this morning.","voice":"samar","timestamps":"sentence"}'API preview. Details may change before launch.
Make some noise.
Early access opens soon. Leave your email and we'll write once, when it does.