All Models, One API.
All History, One ID.
Jug0 is a high-performance LLM gateway built with Rust. It unifies access to top models and provides native, persistent memory for conversations. Stop managing state, start building.
The Developer-First LLM Gateway
Jug0 is packed with features designed to solve real-world problems when building AI applications, saving you time and complexity.
Unified Multi-Provider API
Use GPT-4o, Gemini 1.5, and DeepSeek through a single, consistent API. Swap models with a single line of code, no vendor lock-in.
Native Conversation Memory
Forget manual state management. Just pass the `chat_id` and jug0 automatically retrieves and manages the entire conversation history.
Built with Rust for Speed
Leveraging Axum and Tokio, jug0 delivers extreme performance and low-latency streaming, ready for production-scale workloads.
Streaming-First Architecture
Designed from the ground up for Server-Sent Events (SSE), enabling real-time, typewriter-like experiences in your application effortlessly.
Powerful & Unified Tool Calling
Jug0 normalizes the differences between OpenAI and Gemini's tool calling APIs, providing a single, reliable interface to build powerful agents.
Deploy in Seconds
Packaged as a lightweight Docker container. Get your production-ready LLM backend running with a single command.
Simple Logic, Powerful Results
Jug0 follows a clear, three-stage process for every request, ensuring performance, data integrity, and a seamless developer experience.
1. Ingest & Identify
Your app sends a request. If a `chat_id` is present, jug0 identifies it as an ongoing conversation and retrieves its history from the database.
2. Route & Augment
Jug0 selects the correct LLM provider (e.g., OpenAI, Google) and intelligently augments the request with the full conversation history.
3. Stream & Persist
Jug0 streams the LLM's response back to your app in real-time while simultaneously persisting the new messages to the database for future context.