Crate lc_providers 

LLM provider integrations for langchainrust.

This crate provides unified chat model interfaces for:

• OpenAI (GPT-5 / GPT-4o / o-series)
• Generic OpenAI-compatible endpoints (openai_compatible): Groq, OpenRouter, xAI (Grok), and arbitrary self-hosted/private base URLs
• Ollama (local LLMs like Llama, Mistral)
• DeepSeek (cost-effective Chinese LLM)
• Moonshot (long-context Kimi)
• Qwen (Alibaba Cloud)
• Zhipu (ChatGLM)
• Anthropic (Claude)
• Gemini (Google)

Re-exports

pub use client::LLMClient;

pub use ollama::OllamaChat;

pub use ollama::OllamaConfig;

pub use openai::AssistantError;

pub use openai::BuiltinTool;

pub use openai::OpenAIAssistant;

pub use openai::OpenAIChat;

pub use openai::OpenAIConfig;

pub use openai::ResponsesConfig;

pub use openai::ResponsesError;

pub use openai::ResponsesModel;

pub use openai::OPENAI_MODELS;

pub use openai_compatible::GroqChat;

pub use openai_compatible::OpenAICompatibleChat;

pub use openai_compatible::OpenAICompatibleConfig;

pub use openai_compatible::OpenRouterChat;

pub use openai_compatible::XaiChat;

pub use openai_compatible::DEFAULT_GROQ_MODEL;

pub use openai_compatible::DEFAULT_XAI_MODEL;

pub use openai_compatible::GROQ_BASE_URL;

pub use openai_compatible::GROQ_MODELS;

pub use openai_compatible::OPENROUTER_BASE_URL;

pub use openai_compatible::XAI_BASE_URL;

pub use openai_compatible::XAI_MODELS;

pub use providers::AnthropicChat;

pub use providers::AnthropicConfig;

pub use providers::AnthropicError;

pub use providers::AnthropicStreamToken;

pub use providers::AnthropicStructuredOutputMethod;

pub use providers::AnthropicUsage;

pub use providers::AzureOpenAIChat;

pub use providers::AzureOpenAIConfig;

pub use providers::AzureOpenAIError;

pub use providers::CohereChat;

pub use providers::CohereConfig;

pub use providers::CohereError;

pub use providers::DeepSeekChat;

pub use providers::DeepSeekConfig;

pub use providers::GeminiChat;

pub use providers::GeminiConfig;

pub use providers::GeminiError;

pub use providers::GeminiStructuredOutputMethod;

pub use providers::MistralChat;

pub use providers::MistralConfig;

pub use providers::MoonshotChat;

pub use providers::MoonshotConfig;

pub use providers::QwenChat;

pub use providers::QwenConfig;

pub use providers::ThinkingConfig;

pub use providers::ThinkingType;

pub use providers::ZhipuChat;

pub use providers::ZhipuConfig;

Modules

client

LLMClient — zero-config unified entry point. LLMClient — zero-config unified entry point for switching providers

ollama

Ollama local LLM integration. Ollama local model support

openai

OpenAI API integration.

openai_compatible

B5: generic OpenAI-compatible endpoint support (Groq, OpenRouter, xAI, vLLM, LM Studio, private gateways) — one transport, named presets. Generic OpenAI-compatible chat client (B5, v0.22.4).

providers

Third-party provider integrations. Third-party LLM provider integrations.

Structs

ChatModelWrapper

Wrapper that converts any BaseChatModel’s error into ProviderError.

Enums

ProviderError

Unified error type that aggregates all LLM provider errors.

Functions

wrap_chat_model

Wrap any BaseChatModel into an Arc<dyn BaseChatModel<Error = ProviderError>>.