Sign in with Google

Run AI models on your own computer

Download an open model and chat with it on your laptop or server, with no data sent to a cloud.

For anyone who wants to chat with an AI model with no data sent to a cloud. Start with one engine and one chat app. The rest are for faster models or bigger hardware.

10 picks in 3 sections · checked by an editor on Oct 30, 2026 · How we pick

01Start here2 picks

  1. Ollamaollama/ollama

    The easiest start: one command to download and run a model.

    Docker / self-host · Desktop · CLI · MIT · active yesterday · ★ 185k

    94worth
  2. llmfitAlexsJones/llmfit

    Pick this first to see which models your hardware can run.

    CLI · MIT · active yesterday · ★ 39k

    84worth

02Chat apps4 picks

  1. Open WebUIopen-webui/open-webui

    Pick this for a ChatGPT-style web app on top of Ollama.

    Web · Docker / self-host · Custom (read before use) · active 3 days ago · ★ 157k

    87worth
  2. Cherry StudioCherryHQ/cherry-studio

    Pick this for a desktop app that mixes local and cloud models.

    Desktop · AGPL-3.0 · active yesterday · ★ 53k

    86worth
  3. Janjanhq/jan

    Pick this for a simple desktop app that downloads and runs models for you.

    Desktop · Custom (read before use) · active 2 days ago · ★ 45k

    76worth
  4. AnythingLLMMintplex-Labs/anything-llm

    Pick this to chat with your own documents.

    Docker / self-host · Desktop · MIT · active yesterday · ★ 68k

    86worth

03Engines4 picks

  1. llama.cppggml-org/llama.cpp

    The engine under most local tools. Pick it for full control.

    CLI · Library / SDK · MIT · active 22 hours ago · ★ 133k

    90worth
  2. LocalAImudler/LocalAI

    Pick this to replace the OpenAI API with a local server.

    Docker / self-host · MIT · active yesterday · ★ 50k

    87worth
  3. oMLXjundot/omlx

    Pick this on Apple Silicon Macs for the best speed.

    Desktop · Apache-2.0 · active 2 days ago · ★ 23k

    87worth
  4. AirLLMlyogavin/airllm

    Pick this to try very large models on a small GPU, slowly.

    Library / SDK · Apache-2.0 · active 2 days ago · ★ 41k

    83worth

More in AI tools by job

Add-ons that make your coding agent better

Memory, token savers, specs and skills that make Claude Code, Codex and other coding agents faster and cheaper.Inside: Claude-Mem, RTK, Headroom + 7 more10 picks · 6 sections

Open-source AI coding agents to try first

Coding agents you can read, change and run with the model you choose, in the terminal, the editor or the browser.Inside: OpenCode, Codex CLI, Gemini CLI + 5 more8 picks · 4 sections

AI tools for testing web apps and AI features

Let an AI agent click through your web app, write browser tests, or check the answers your own AI feature gives.Inside: Playwright MCP, playwright-cli, Chrome DevTools MCP + 7 more10 picks · 2 sections

Other groups

See every project in this category on Explore →Missing a project? Suggest it →

How we pick

An editor picks 6 to 12 projects for one job and writes when to choose each one. Every pick had a commit in the last 6 months. It is not archived. Its worth score is 60 or more. An editor breaks a rule only for a reason, and writes that reason in the pick's note. We check every list again each quarter. A pick goes when it breaks a rule or a clearly better project appears.