Modules
In-memory ETS-based 3-state circuit breaker.
Core LLM HTTP client for OpenAI-compatible chat completion APIs.
Defensive JSON decoder that tries JsonRemedy.repair/1 first for
malformed LLM output, falling back to Jason.decode/2.
Extracts and parses JSON from LLM responses.
Toggleable structured logger for LLM calls.
In-memory ETS-based metrics collector for LLM calls.
Behaviour for LLM provider implementations.
Registry of all supported LLM providers.
Provider implementation for Fireworks AI API.
Provider implementation for xAI Grok API.
Provider implementation for Groq API.
Provider implementation for Xiaomi Mimo API.
Provider implementation for MiniMax API.
Mock provider for testing purposes.
Provider implementation for NVIDIA NIM API.
Provider implementation for OpenAI API.
Provider implementation for Sarvam AI API.
In-memory ETS-based sliding window rate limiter.
Single entry point for all LLM JSON parsing.