Local LLM and self-hosted AI tools, compared with prices

Updated 23 Oct 2026Prices checked 17 Oct–23 Oct 2026How we check prices

Want AI that runs on your own computer, or on a server you control? Below are the tools in our database that run models locally, then the self-hostable AI tools for teams. Every price comes from the vendor's own page with a checked date and source; open-source licences and GitHub stars come from each repository.

At a glance

Setup difficulty: 1 = install an app, 5 = needs a GPU server and ops skills. Only verified prices are shown.

ToolPriceLicenceSetupGitHub starsChecked
AnythingLLMFrom $50/mo billed monthly · free planMIT167.5k23 Oct 2026 · source
JanFree and open source (self-host)Apache-2.0145.0k22 Oct 2026 · source
OllamaFrom $20/mo billed monthly · free planMIT2183.6k22 Oct 2026 · source
LM StudioFrom $20/mo billed monthly · free planClosed source––22 Oct 2026 · source
MstyFrom $12.42/mo billed annually · free planClosed source––23 Oct 2026 · source
Open WebUIFree and open source (self-host)Open WebUI License3156.0k22 Oct 2026 · source
llama.cppFree and open source (MIT)MIT3132.1k23 Oct 2026 · source
LocalAIFree and open source (MIT), self-hostedMIT349.9k23 Oct 2026 · source
LibreChatFree and open source (self-host)MIT345.9k22 Oct 2026 · source

Run a local LLM on your own computer

Desktop apps and model runners that download open models and run them on your machine. Easiest first.

  1. AnythingLLM

    Local AI

    Best for: People and small teams who want private document chat on their own machine or server

    Open-source desktop and Docker app to chat with your documents and run AI agents with local or cloud models; the desktop app needs no account or API key.

    Free planOpen source · MITSelf-hostable

    From $50/mo billed monthly · free plan

    Checked 23 Oct 2026 · source (AnythingLLM pricing source, opens in a new tab)

    All plans

    Desktop app and Docker self-host free (MIT); hosted Basic $50/mo; Pro $99/mo; Enterprise custom

    ★ 67.5k · MIT · Setup 1/5 · repo (AnythingLLM GitHub repository)

  2. Ollama

    Local AI

    Best for: Developers running models on their own machine

    Run open-weight LLMs locally from the command line.

    Free planOpen source · MITSelf-hostable

    From $20/mo billed monthly · free plan

    Checked 22 Oct 2026 · source (Ollama pricing source, opens in a new tab)

    All plans

    Free (run models locally); Pro $20/mo or $200/yr ($16.67/mo billed annually); Max $100/mo; Team $500/mo (early access)

    ★ 183.6k · MIT · Setup 2/5 · repo (Ollama GitHub repository)

  3. Msty

    Local AI

    Best for: People who want one app for local and cloud models without the command line

    Desktop AI workspace (Msty Studio) to chat with local and online models side by side, with knowledge stacks and an agent mode.

    Free plan

    From $12.42/mo billed annually · free plan

    Checked 23 Oct 2026 · source (Msty pricing source, opens in a new tab)

    All plans

    Free $0 forever (Studio Desktop core features); Aurum $149/user/yr; Aurum Lifetime $349/user one-time

    Starting price $12.42/mo is billed annually

  4. llama.cpp

    Local AI

    Best for: Technical users who want the lowest-level way to run open models locally

    Open-source LLM inference in C/C++, with a command-line tool and llama-server for running open models on your own machine.

    Free planOpen source · MITSelf-hostable

    Free and open source (MIT)

    Checked 23 Oct 2026 · source (llama.cpp pricing source, opens in a new tab)

    All plans

    Free and open source (MIT)

    ★ 132.1k · MIT · Setup 3/5 · repo (llama.cpp GitHub repository)

  5. LocalAI

    Local AI

    Best for: Developers who want a drop-in local replacement for a cloud AI API

    Open-source, OpenAI-compatible inference server that runs text, voice, vision and image models on the hardware you already own.

    Free planOpen source · MITSelf-hostable

    Free and open source (MIT), self-hosted

    Checked 23 Oct 2026 · source (LocalAI pricing source, opens in a new tab)

    All plans

    Free and open source (MIT), self-hosted

    ★ 49.9k · MIT · Setup 3/5 · repo (LocalAI GitHub repository)

Self-hosted AI tools for your own server

Open-source or source-available AI tools you can run on your own server, grouped by category. Each card shows its licence and price.

  1. OpenHands

    AI coding agent

    Best for: Teams wanting an open-source Devin-style agent

    Open-source autonomous software-engineering agent you can run locally or self-host, with paid self-hosted Teams and Enterprise plans.

    Free planOpen source · MITSelf-hostable

    From $499/mo billed monthly · free plan

    Checked 22 Oct 2026 · source (OpenHands pricing source, opens in a new tab)

    All plans

    Agent Canvas free (open source, self-hosted); Teams $499/mo for up to 10 users (+$50/mo per extra seat); Enterprise custom pricing

    ★ 91.1k · MIT · Setup 3/5 · repo (OpenHands GitHub repository)

  2. Stable Diffusion WebUI (AUTOMATIC1111)

    AI image generation

    Best for: Hobbyists with a GPU

    Popular open-source web UI for running Stable Diffusion locally.

    Free planOpen source · AGPL-3.0Self-hostable

    Free and open source (self-host)

    Checked 22 Oct 2026 · source (Stable Diffusion WebUI (AUTOMATIC1111) pricing source, opens in a new tab)

    All plans

    Free and open source (self-host)

    ★ 165.1k · AGPL-3.0 · Setup 4/5 · repo (Stable Diffusion WebUI (AUTOMATIC1111) GitHub repository)

  3. VoiceStudio

    AI voice

    Best for: Creators who want ElevenLabs-style cloning and dubbing on their own machine

    Open-source, fully local voice cloning, voice design, video dubbing, dictation, transcription and audiobook creation in 646 languages.

    Free planOpen source · AGPL-3.0Self-hostable

    From $8.25/mo billed annually · free plan

    Checked 22 Oct 2026 · source (VoiceStudio pricing source, opens in a new tab)

    All plans

    Free and open source (AGPL-3.0); Pro $99/user/yr; Lifetime $299 one-time

    Starting price $8.25/mo is billed annually

    ★ 79.2k · AGPL-3.0 · repo (VoiceStudio GitHub repository)

  4. LanguageTool

    AI writing assistant

    Best for: Multilingual writers

    Multilingual grammar and style checker, with an open-source core.

    Free planOpen source · LGPL-2.1Self-hostable

    India price (INR) · USD price not available

    Premium ₹999/mo monthly or ₹3,499/yr (India pricing) · free plan

    Checked 22 Oct 2026 · source (LanguageTool pricing source, opens in a new tab)

    All plans

    Free plan; Premium ₹999/mo monthly or ₹3,499/yr (India pricing)

    ★ 15.2k · LGPL-2.1 · Setup 3/5 · repo (LanguageTool GitHub repository)

  5. n8n

    Automation

    Best for: Technical teams wanting control and self-hosting

    Fair-code workflow automation with AI agents; self-host or cloud.

    Free planFair-code (source-available)Self-hostable

    Price in EUR · USD price not available

    Cloud Starter €20/mo, Pro €50/mo (billed annually) · free plan

    Checked 22 Oct 2026 · source (n8n pricing source, opens in a new tab)

    All plans

    Self-hosted Community Edition free; Cloud Starter €20/mo, Pro €50/mo (billed annually)

    ★ 207.8k · Sustainable Use License · Setup 3/5 · repo (n8n GitHub repository)

Price not yet verified on the vendor's own page in the last 14 days, so not ranked:

Related pages

FAQ

What is a local LLM?

A large language model that runs on your own computer or server instead of a vendor's cloud, so prompts and documents stay on your hardware. Apps like the ones on this page download open models and run them for you.

Which local LLM apps are free?

On this page: AnythingLLM, Jan, Ollama, LM Studio, Msty, Open WebUI, llama.cpp, LocalAI, LibreChat have a free version or free plan, checked against each vendor's own page.

Which local AI tools are open source?

AnythingLLM, Jan, Ollama, llama.cpp, LocalAI, LibreChat. Licence data comes from each project's GitHub repository.

What is the easiest way to start?

Install a desktop app (setup difficulty 1 in the table) and download a small model inside it. Server tools such as LocalAI and llama.cpp need the command line.

Can I use a local model for coding?

Several coding agents let you connect your own model provider. See our Claude Code and Cursor alternatives pages for which ones, with prices.

Get the weekly AI price changes email

One email a week: AI tool price changes and new alternatives, with sources. Free, no spam, unsubscribe any time. See our privacy policy.