What problem does it solve?
It helps you correctly integrate Sarvam AI’s Indian-language APIs so your agents can chat, transcribe, translate, synthesize speech, detect languages, and process documents without getting stuck on authentication, model quirks, or mismatched SDK methods.
Core Features & Use Cases
- Chat LLM integration: Generate multilingual responses using the Sarvam chat completions endpoint with the correct headers and SDK call patterns.
- Speech & language pipelines: Convert speech to text (with required
mode for Saaras v3), translate speech to English text, synthesize speech from text, and handle transliteration and language detection.
- Document intelligence: Run the Sarvam Vision batch pipeline to extract and retrieve results from documents via the required multi-step flow.
Quick Start
Ask an AI to generate a working example request for Sarvam chat completions using model sarvam-105b with Bearer authentication and return a response with correct handling for reasoning_effort-related fields.