Tabby is the open-source, self-hosted AI coding assistant that lets you keep your code where it belongs - on your own machines, not floating around someone else's cloud. Think of it as GitHub Copilot's privacy-obsessed cousin who insists on running everything locally, no external databases required. Drop it in with a single Docker command and you're off to the races.

Main Features

• Self-hosted code completion: Full AI-powered inline code suggestions running entirely on your own infrastructure, with no data ever leaving your servers.
• Answer Engine: A built-in chat interface that answers coding questions by pulling context from your codebase, documentation, and internal knowledge - right inside your IDE.
• No external dependencies: Completely self-contained with no need for a DBMS or cloud service - just Docker and a GPU (consumer-grade works fine).
• Multi-IDE support: Extensions available for VS Code, JetBrains IDEs, and Vim/Neovim, with an OpenAPI interface for integrating into custom or cloud IDEs.
• Flexible model selection: Swap between different backend models including StarCoder, CodeLlama, CodeQwen, Codestral, and more - pick what works for your stack.
• Repository context awareness: Uses RAG-based code completion that pulls relevant snippets from your entire repo, including recently modified files and LSP declarations.
• Team management and analytics: Built-in admin UI with team management, usage analytics, and activity tracking so you know who's using what.
• Enterprise authentication: LDAP integration, GitLab SSO, GitHub/GitLab context indexing, and secured access controls for larger teams.

Who Should Use It?

• Privacy-conscious developers: Anyone who wants AI code completion without shipping their proprietary codebase to a third-party cloud.
• Enterprise engineering teams: Companies that need AI coding tools but have strict data residency or compliance requirements (SOC 2, HIPAA, etc.).
• Startup engineering leads: Teams that want Copilot-level assistance at a fraction of the per-seat cost by running their own instance.
• Open-source maintainers: Developers who prefer open-source tooling and want to inspect, modify, or contribute to their AI coding assistant.
• Cloud IDE operators: Platforms running browser-based IDEs that want to plug in AI code completion via Tabby's OpenAPI interface.
• Freelance developers: Solo devs who can run Tabby on their own workstation GPU and avoid monthly subscription fees.