01 — Malayalam voice infrastructure

നാട്ടുമൊഴി

/ the native tongue

Voice your app can actually talk with, in Malayalam.

ker-ai is a hosted Malayalam voice API — speech in, tool calls against your own functions, and spoken replies. Built for one language, done exceptionally well.

Request access Read the docs

02 — The API

Stream voice over the API.

// /ws/voice

// client -> ker-ai
{ "type": "session.init", "payload": { "language": "ml-IN" } }
{ "type": "audio.chunk",  "payload": { "base64": "..." } }

// ker-ai -> client
{ "type": "transcript.final", "payload": { "text": "..." } }
{ "type": "function.call",     "payload": { "name": "...", "arguments": {} } }
{ "type": "assistant.audio",   "payload": { "base64": "..." } }

Speech becomes action

A user speaks; ker-ai calls the functions your app already exposes and speaks the result back. Real in-app actions, not a chatbot bolted on the side.

Malayalam, exactly

One underserved language done well beats ten done passably. Malayalam is the proof the whole service is built around, never a dropdown afterthought.

A service you call, not an app

ker-ai is hosted infrastructure you reach over an API — it runs the speech and the reasoning behind your product. Your users never download a separate assistant; they just talk to you.