Tabby is a self-hosted AI coding assistant built for teams that can't send their source code to a third-party cloud. You run it on your own hardware — including your own GPU — and it delivers context-aware completions across popular IDEs while your code never leaves your network.
It indexes your repositories to give suggestions grounded in your actual codebase, supports team deployment with admin controls, and works with a range of open models. That makes it a genuine on-prem alternative to GitHub Copilot for privacy-sensitive organisations.
It's Apache-2.0 licensed with an active community and steady releases. For regulated industries or any team that wants Copilot-style help without the data leaving the building, Tabby is the leading open, self-hosted choice.
Run Tabby on your own hardware and GPU — your source code never leaves your network.
It indexes your repositories to ground suggestions in your actual codebase.
Admin controls and multi-user support make it practical for whole engineering teams.
Works across popular editors so developers get help where they already work.
Yes — Tabby is free and open source under the Apache-2.0 license. You provide the hardware to run it.
Tabby is self-hosted and open source, so your code stays on your own infrastructure, while Copilot is a closed cloud service.
A GPU is recommended for the best completion speed and quality, though smaller setups are possible.
Tabby provides extensions for a range of popular IDEs, so developers get completions in their usual editor.
Yes — because it's fully self-hosted, it suits teams with strict privacy, compliance or data-residency requirements.
Open-source AI code assistant for VS Code and JetBrains.
Agentic AI coding assistant for VS Code with multi-file edits.
Open-source AI coding agent for the terminal with 75+ LLM providers.
Claim Tabby to add screenshots, respond to reviews and reach founders comparing tools — free to start.