DEEIX-AI/DEEIX-Chat

What it solves

DEEIX Chat provides a unified, enterprise-grade platform for accessing multiple AI model providers through a single entry point. It eliminates the complexity of managing various upstream channels, billing, and identity systems by integrating them into one deployable product for individuals, teams, and enterprises.

How it works

The system uses a split architecture with a Next.js/React frontend and a Go-based backend runtime. The backend handles API orchestration, authentication, and model routing, while the frontend serves as the user interface. It supports multiple database options (PostgreSQL with pgvector or SQLite with sqlite-vec) and storage solutions (Local or S3), allowing for flexible deployment from lightweight single-node setups to high-concurrency production environments.

Who it’s for

It is designed for individuals, teams, and enterprises that require stable, long-term access to various LLM providers and need integrated operational controls like usage billing, audit logs, and identity management.

Highlights

• Model Routing: A platform-model layer that manages upstream channels, priority, weights, and circuit breaking to reduce multi-provider operational costs.
• Multimodal Capabilities: Support for text, image, and tool-based interactions across providers like OpenAI, Anthropic, Google/Gemini, xAI, and OpenRouter.
• Integrated RAG: Built-in file upload, OCR, chunking, embeddings, and semantic retrieval to inject document content into conversations.
• Enterprise Operations: Comprehensive tools for usage billing (including Stripe and EPay), 2FA/TOTP security, SSO/OIDC/OAuth identity, and detailed audit logs.
• Tool Ecosystem: Integration with MCP servers and provider-native tools with full traceability and execution limits.