Skip to content

MindsHub

An open-source agent workspace for knowledge work and software, with a choice of models.

Stars
39.8k
Forks
6.2k
Licence
MIT
Language
Makefile
Last update
13 days ago
Type
App

What it is

MindsHub is an agent workspace where AI agents help with research, analysis, content, and software development. It lets you pick among many models and providers rather than being tied to one. The project can run as a desktop app, in a browser, or be built from source on your own infrastructure.

  • Automate a recurring report that reads from connected data sources and writes a summary
  • Turn agent output into a dashboard, document, or small app and publish it to a live URL
  • Connect data sources like Postgres, Gmail, or Notion through its vault and let agents query them with scoped credentials
  • Switch between Claude, GPT, Gemini, DeepSeek, or Qwen models from one interface

Great for

  • Knowledge workers who want to automate reading and writing tasks like reports and recurring workflows
  • Developers who want to build and run the full MindsHub stack from source
  • Teams that need to deploy in a cloud, VPC, on-prem, or air-gapped setup
  • People who want to compare or switch between frontier and open models

Not ideal for

  • Users who want a fully offline, minimal tool: the README describes connections to hosted providers and a paid Pro tier, and does not detail offline operation
  • Anyone needing a stable, fully documented self-hosting path: the README is brief on build internals and links out to docs for setup
  • Linux users who want a packaged installer: the README points Linux users to building from source
Step-by-step guide: how to use MindsHubInstall, first run, examples, tips and FAQs

How to get it

  1. 1

    Clone and build from source

    You'll need: Git · Make · Python 3.10–3.13 (per README badge)

    Clone the repository with submodules

    git clone --recurse-submodules https://github.com/mindsdb/minds.git

    Enter the project directory

    cd minds

    Install dependencies

    make setup

    Run the desktop app with hot reload

    make dev
  2. 2

    Run it

    • Sign in at console.mindshub.ai for the zero-install option, or download the macOS or Windows desktop app
    • For source builds, clone the repo with submodules and run make setup
    • Start the desktop app with make dev, or the browser version with make dev-web
    • Choose a model and open agent from the dropdown and begin a task
  3. 3

    Get the most out of it

    • Use the Connected data vault to link systems like BigQuery, Postgres, Gmail, Drive, HubSpot, Notion, and Linear with scoped credentials
    • Schedule recurring tasks and use the reusable skill library and cross-session memory
    • Publish artifacts such as dashboards and apps to a live URL to share with your team
    • Set feature branch refs in a gitignored dev.env and use make use and make refs to control which branches run

Downloads

Strengths

  • Model choice: switch between frontier and open models without a separate key for each provider, per the README
  • Open-source agent harnesses (Anton and Hermes) with the project released under the MIT license
  • Multiple run options: web app, macOS and Windows desktop apps, or build from source

Trade-offs

  • Pro tier is needed for all frontier models and private artifacts, according to the README
  • The README is truncated and light on technical detail, so some behavior is unclear from it alone
  • Building from source requires cloning with submodules and running make targets; `make flush` deletes conversations and saved keys if run

How it compares

ProjectStarsTypeLevelUpdatedBest for
mindshub (this)39.8kAppintermediate13 days agoKnowledge workers who want to automate reading and writing tasks like reports and recurring workflows
LibreChat46.6kAppintermediate4 days agoTeams that want one shared chat interface across several AI providers
anything-llm67.8kAppbeginner6 days agoPeople who want a private ChatGPT-style assistant over their own PDFs and text files
OmniRoute83.8kAppintermediate4 days agoDevelopers who use several AI coding tools and want one place to configure providers
hermes-agent260kAI model / agentintermediate4 days agoDevelopers who want a personal agent they can host on their own server or machine
CowAgent47.8kAppintermediate4 days agoPeople who want a self-hosted personal assistant reachable from their everyday chat apps

Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.