nexo-llm

LLM provider clients (MiniMax, OpenAI-compat, Anthropic, Gemini) with rate limiter and tool registry.

This crate is part of Nexo — a multi-agent Rust framework with a NATS event bus, pluggable LLM providers (MiniMax, Anthropic, OpenAI-compat, Gemini, DeepSeek), per-agent credentials, MCP support, and channel plugins for WhatsApp, Telegram, Email, and Browser (CDP).

• Main project: https://github.com/lordmacu/nexo-rs
• Documentation: https://lordmacu.github.io/nexo-rs/

What this crate does

• Pluggable LlmClient trait — swap providers without touching the agent runtime.
• Built-in providers: MiniMax M2.5 (primary), OpenAI-compatible (any OpenAI-format endpoint), Anthropic Claude (API key and subscription/OAuth flow), Gemini, DeepSeek.
• Tool registry with JSON-schema validation, per-binding tool gating, parallel tool calls.
• Rate limiter + retry wired through nexo-resilience (LLM 429 → 5 attempts 1s→60s exp, 5xx → 3 attempts 2s→30s).
• Streaming responses, prompt caching helpers (Anthropic), and reasoning-mode passthrough.

Install

[dependencies]
nexo-llm = "0.1"

Documentation for this crate

• MiniMax M2.5
• Anthropic / Claude
• OpenAI-compatible
• DeepSeek
• Rate limiting & retry
• llm.yaml

License

Licensed under either of:

• Apache License, Version 2.0 (LICENSE-APACHE)
• MIT license (LICENSE-MIT)

at your option.