API Reference

Modules

Module Description
Chat Chat module - Stateful multi-turn conversation sessions
Tokenizer Tokenizer module - Text encoding and decoding
Template User-defined prompt template engine for LLM applications
Validate JSON schema validation with streaming support
Db Storage and persistence for conversations and vectors

Utilities

Module Description
Converter Model Conversion API
Repository Model repository management

Supporting

Module Description
Client Entry point for LLM inference
Profile Profile-based chat session persistence for Python users
Types Open Responses data model
Router Router module - Model routing, backend specification, and generation config
Exceptions Talu exceptions

Convenience Functions

Function Description
talu.ask() One-shot question-answer with automatic resource cleanup.
talu.list_sessions() List sessions in a profile.
talu.convert() Convert a model to an optimized format for efficient infe...