agentix 0.25.0

Multi-provider LLM client for Rust — streaming, non-streaming, tool calls, MCP, DeepSeek, OpenAI, Anthropic, Gemini, Mimo
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38

//! HTTP servers that expose agentix as an LLM-API-compatible endpoint.
//!
//! Following the pandoc model: agentix's internal `Request` + `Message` +
//! `LlmEvent` AST is the hub; the existing 10 outbound providers are
//! "writers"; the modules here are "readers" that accept inbound requests in
//! various wire formats and translate them to the AST.
//!
//! Currently implemented:
//!
//! - [`anthropic`] — Anthropic Messages format on `POST /v1/messages`.
//! - [`openai_chat`] — OpenAI Chat Completions on `POST /v1/chat/completions`
//!   (gated by the `server-openai-chat` feature).
//!
//! All readers share one fallback chain (see [`UpstreamSpec`]) and one shared
//! state, so a single bind port can serve multiple wire formats simultaneously
//! by merging their routers — see the `agentix` CLI binary for an example.

pub mod fallback;
pub mod translated;

pub mod anthropic;

#[cfg(feature = "server-openai-chat")]
pub mod openai_chat;

#[cfg(feature = "server-openai-responses")]
pub mod openai_responses;

pub use anthropic::AnthropicServer;
pub use anthropic::error::ServerError;
pub use fallback::UpstreamSpec;
pub use translated::Translated;

#[cfg(feature = "server-openai-chat")]
pub use openai_chat::OpenAIChatServer;

#[cfg(feature = "server-openai-responses")]
pub use openai_responses::OpenAIResponsesServer;