Tabby

Self-hosted coding assistant with a local inference server.

Install Repository Website Compare

Fit.

Catalog snapshot

Platforms: Linux, Macos, Windows, Docker
Operating systems: Linux, macOS, Windows
Ram floor: 16 GB
Gpu: Optional · 8 GB VRAM helps
Docker: Yes
Setup: Advanced
Deploy as: Local, Self-hosted
Commercial use: Allowed

CPU works for tiny models. A GPU is the honest setup.

Project.

Fetched Sep 12, 2026¹

License: Apache 2.0
Stars: 33.9k¹
Forks: 1.8k¹
Open issues: 261¹
Language: Rust¹
Last commit: Jun 30, 2026¹
Latest release: Aug 12, 2026
Tags: self-hostedautocompletegpu

About.

Tabby is Copilot-like completion you host. It ships a server, a VS Code client, and a path to run models on your GPU. Heavier than Continue-plus-Ollama.

completion server · IDE clients · admin UI

Alternatives.

Nearby.

All Coding agents

GCGemini CLIGoogle’s Apache-2.0 terminal coding agent with MCP tools and a large free tier.Compare
GgptmeMIT terminal agent toolkit with code, shell, and web tools — local-model friendly.Compare
OOpenCodeOpen-source coding agent with a terminal UI, built to run against local or hosted models.Compare
OOpenHandsSoftware agent that uses a sandbox to edit code, run tests, and open PRs.Compare
GGooseBlock’s local-first agent runtime with a desktop and CLI.Compare