Tabby
Self-hosted AI coding assistant, an open-source on-premises alternative to GitHub Copilot
Install
docker run -it \
--gpus all -p 8080:8080 -v $HOME/.tabby:/data \
tabbyml/tabby \
serve --model StarCoder-1B --device cuda --chat-model Qwen2-1.5B-InstructRun the server in Docker with GPU
This is third-party code. Review the repository files before installing.
What it does
Tabby runs on your own server and provides code completion and chat in the IDE without cloud services or a separate DBMS. The server runs on consumer-grade GPUs, exposes an OpenAPI interface for integration and has extensions for VS Code, JetBrains and Vim. It includes an Answer Engine for internal documentation, repository and GitLab Merge Request indexing as context, team management and usage reports.
Who it is for. Teams that cannot send code to external cloud services.
Good fit when
- You need code completion inside a closed network
- You want to run open models on your own GPU
- You need answers from internal documentation and repositories
- You need LDAP auth and team usage reports
Not a fit when
- You have no GPU server and do not want to maintain one
- You need an autonomous agent that runs multi-step tasks on its own
Example request
Deploy Tabby on our GPU server in Docker with the StarCoder-1B model and connect the VS Code extensionLimitations
Quality depends on the chosen models and hardware. The README quick start assumes Docker with an NVIDIA GPU. Agent features live in a separate product, Pochi.
How to disable. Stop the Tabby container and remove the Tabby extension from your IDE.
Security check
- Runs a server and downloads models
- Indexes company repositories and documentation
README in short
Tabby is a self-hosted AI coding assistant and an open-source alternative to GitHub Copilot. It needs no DBMS or cloud, exposes an OpenAPI interface and runs on consumer GPUs. The README lists release history with features such as Answer Engine, GitLab MR indexing and LDAP. It gives a Docker run command and instructions for building from Rust source.
FAQ
Does it need a database or cloud?
No, according to the README Tabby is self-contained and needs no DBMS or cloud service.
Related
A skills library that gives coding agents a development process: brainstorming, planning, TDD, subagents and code review
Skills for real engineers by Matt Pocock
Skills For Real Engineers
Small composable skills for engineering with agents: plan grilling, TDD, bug diagnosis, code review and architecture
Reference MCP servers
Model Context Protocol servers
Official reference MCP servers: Filesystem, Fetch, Git, Memory, Sequential Thinking, Time and Everything
Up-to-date, version-specific library docs and code examples in your agent's context, via MCP or a CLI plus skill