Practical overview
When it fits
Use Deeix Chat when you need to operate a model-provider workspace yourself and want chat, routing, retrieval, tools and usage administration in the same application. Choose the SQLite setup for lightweight single-process use; the documentation points multi-node or high-concurrency deployments toward PostgreSQL and Redis.
Who it is for
- Teams and enterprises seeking a shared workspace across multiple model providers.
- Operators who want to configure model routes, provider connections, usage controls and user administration.
- Developers evaluating a self-hosted AI chat platform with file retrieval and tool integration.
Documented uses
- Route requests to multiple upstream model providers through one workspace.
- Chat with models using text, images and files, and bring document content into context through retrieval.
- Enable MCP or provider-native tools and inspect tool-call traces.
- Manage model routes, pricing, balances, subscriptions, usage records and audit events.
- Deploy a chat and administration application using SQLite for a single-process setup or PostgreSQL and Redis for a larger deployment.
Requirements & limits
The project is open source under the Apache-2.0 license and documents Docker-based self-hosting.
- Model routing includes route bindings, priority, weights, circuit breaking and provider capability configuration.
- Supported provider protocols include OpenAI, Anthropic, Google/Gemini, xAI, OpenRouter and OpenAI-compatible protocols.
- File features range from upload and extraction to chunking, embeddings and semantic retrieval; heavy extraction and OCR services are optional.
- SQLite with an in-process cache is single-process only. PostgreSQL with Redis is the documented choice for multi-node or high-concurrency deployments.
- The frontend is built into static assets served by the Go backend runtime; the application is not just a model API gateway.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Start a lightweight SQLite deployment
Launch the documented single-container SQLite setup for local evaluation or a small single-node deployment.
Before you start
- Docker Compose
- The Deeix Chat repository's deploy directory
Steps
Change to the deploy directory.
Copy config.sqlite.example.yaml to config.yaml.
Start the SQLite configuration with docker compose -f docker-compose.sqlite.yml up -d.
Expected outcome
The app container runs with SQLite, local vector indexes and an in-process memory cache.
Guide limitations
- SQLite with memory cache is single-process only; the README recommends PostgreSQL and Redis for multi-node or high-concurrency production deployments.
Prepare the full Docker Compose stack
Configure the documented compose deployment that starts the app, PostgreSQL and Redis together.
Before you start
- Docker Compose
- The Deeix Chat repository's deploy directory
Steps
Change to the deploy directory.
Copy config.full.example.yaml to config.yaml.
Start the full stack with docker compose -f docker-compose.full.yml up -d.
Expected outcome
Compose starts the application together with PostgreSQL and Redis; the compose environment variables for those services override corresponding values in config.yaml.
Guide limitations
- Review and configure infrastructure settings before deployment; environment variables take precedence over config.yaml.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 8 October 2026
- Original website title
- DEEIX-AI/DEEIX-Chat · Official homepage title · retained in its source language