DOCSEARCHbyAlgolia
Submit Docs

MindsHub (Cowork)

repository·main·Indexed 13 days ago

https://github.com/mindsdb/mindshub

A unified workspace for delegating complex AI projects—such as app building, research, and automated reporting—to open-source and proprietary models. It integrates a desktop/web app, an agent backend, and a data engine, featuring model routing for Claude, GPT, Gemini, DeepSeek, Qwen, and Kimi, as well as interchangeable agent harnesses like Anton and Hermes.

Tokens
4.6K
Snippets
15
Records
20
Agent score
93%

What's inside MindsHub

  1. What is MindsHub Cowork?

    main

    MindsHub Cowork is an integrated workspace designed to delegate entire projects—such as apps, websites, research, analysis, reporting, and scheduled operations—to AI models.

    Key capabilities include:

    • Connecting Data: Securely connect to systems like BigQuery, Postgres, Gmail, Drive, HubSpot, Notion, and Linear. Credentials are scoped per connection, ensuring agents never see raw keys.
    • Model Routing: Switch between frontier models (Claude, GPT, Gemini) and open models (DeepSeek, Qwen, Kimi) without managing individual provider keys.
    • Open Agents: Run interchangeable open-source harnesses like Anton (default) and Hermes via a dropdown menu.
    • Artifacts: Transform agent outputs into documents, dashboards, apps, and code, which can be published to a live URL.
    • Memory, Skills, and Scheduling: Includes session-persistent memory, reusable skill libraries, and scheduled task execution.

  2. Core features of MindsHub

    main

    MindsHub provides several key capabilities for AI-driven workflows:

    • Data Connections: A secure vault to connect systems like BigQuery, Postgres, Gmail, Drive, HubSpot, Notion, and Linear. Credentials are scoped to individual connections so agents never see raw keys.
    • Model Router: Switch between frontier models (Claude, GPT, Gemini) and open-source models (DeepSeek, Qwen, Kimi) without configuring individual provider keys.
    • Open Source Agents: Run interchangeable execution engines (harnesses) like Anton (default) or Hermes via a dropdown menu.
    • Artifacts: Convert agent outputs into documents, dashboards, applications, or code, which can be published via online links.
    • Memory, Skills, and Scheduled Tasks: Includes cross-session memory, reusable skill libraries, and scheduled task execution.

  3. How to work on feature branches with submodules

    main

    MindsHub is a superproject that pins specific commits for each module (frontend, backend/core_api, backend/core_agent, backend/data-vault). To work on a specific feature branch without affecting the superproject's git status or pins, follow this workflow:

    1. Configure your branch

    Copy the template to dev.env and set the REF variable to your desired feature branch (e.g., feat/my-thing).

    cp dev.env.example dev.env
# Edit dev.env to set REF=feat/my-thing

    2. Use make to manage submodules

    The make system uses your dev.env settings to manage submodule checkouts and application runs.

    cp dev.env.example dev.env
# Set REF=feat/my-thing in dev.env

  4. Build MindsHub from source

    main

    To build the full stack (desktop/web app, agent backend, and data engine) from source, follow these steps:

    1. Clone the repository with submodules:

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

    2. Install dependencies:

      make setup

    3. Run the application: Use the following commands depending on your target environment:

      • Desktop app (Electron) with hot reload: make dev or make watch
      • Web app in browser with hot reload: make dev-web
      • Production build: make build
      • Package for macOS: make dist-mac
      • Package for Windows: make dist-win
      • Build macOS .app from local uncommitted source: make pack-local

    Resetting the installation

    To perform a fresh start (deleting local runtime, virtual environments, and app state in ~/.anton and ~/.cowork), use:

    make flush

    Warning: This deletes all conversations and saved keys. Use FORCE=1 to skip the confirmation prompt.

    git clone --recurse-submodules https://github.com/mindsdb/minds.git
cd minds
make setup
make dev

  5. Get started with MindsHub Cowork

    main

    MindsHub Cowork provides several ways to access the platform depending on your needs:

    • Web: No installation required. Access via console.mindshub.ai.
    • macOS: Download the .pkg desktop application.
    • Windows: Download the .exe desktop application.
    • Linux: Requires building from source.

    Free tier is available; the Pro plan includes advanced models and private artifacts.

  6. Get started with MindsHub

    main

    MindsHub (Cowork) is a unified workspace for delegating complete projects like apps, research, and automated workflows to AI models. You can use it via the web, desktop applications, or by building from source.

    Quick Start Options

  7. Develop on feature branches using submodules

    main

    MindsHub is a superproject where modules (frontend, backend/core_api, backend/core_agent, backend/data-vault) are pinned to specific commits. To work on feature branches without affecting the superproject's git status, use a dev.env file.

    1. Configure your branches: Copy the template and define your target branches (e.g., REF=feat/my-feature or API_REF=feat/api-feature):

    cp dev.env.example dev.env

    2. Use make commands to manage submodules:

    CommandDescription
    make useChecks out the branches defined in dev.env across all submodules
    make dev / make watchRuns Electron app with hot reload against local code
    make dev-webRuns the web SPA with hot reload against local code
    make server + make appReinstalls and launches the desktop server from the configured branch
    make server-local + make app-localInstalls and launches the desktop server from local code without confirmation
    make pack-localCompiles macOS .app from local code without needing a push
    make refsDisplays which references will be used in the next execution
    make baselineResets submodules to the pinned commits
    make pinRegisters current submodule commits as the new pins for the superproject

    Because submodules are configured with ignore = all, your branch work will not appear as changes in the parent repository's git status.

    cp dev.env.example dev.env
# Edit dev.env to set REF or API_REF
make use

  8. Develop on MindsHub feature branches (submodules)

    main

    MindsHub is a superproject that pins modules (frontend, backend/core_api, backend/core_agent, backend/data-vault) to specific commits. To work on feature branches without affecting the superproject's pinned state, use a dev.env file.

    Workflow

    1. Configure branches: Copy the template and set your desired branch references (e.g., REF=feat/my-thing or API_REF=...).

      cp dev.env.example dev.env

    2. Use make commands to manage submodules:

      • make use: Check out the branches specified in your dev.env across all submodules.
      • make refs: Show which references will be used in the next run.
      • make baseline: Reset submodules to the original pinned commits.
      • make pin: Record the current submodule commits as the new superproject pins (requires a deliberate commit).

    Running development environments

    • make dev / make watch: Run Electron app with live reload.
    • make dev-web: Run web SPA with live reload.
    • make server + make app: Reinstall desktop server from configured branch, then launch.
    • make server-local + make app-local: Install desktop server from local uncommitted source, then launch.
    cp dev.env.example dev.env
# Edit dev.env to set REF=your-branch
make use
make dev

  9. Configure per-developer module branches via dev.env

    main

    MindsHub allows developers to specify which git branches to use for different modules during development. To configure this, copy dev.env.example to a new file named dev.env (which is gitignored).

    Once configured, you can use the following commands:

    • make use && make dev: Checks out the specified branches and starts local-source development.
    • make server && make app: Installs and runs the matching Electron desktop server for your branches.

    If no variables are set, the system defaults to tracking the main branch.

    cp dev.env.example dev.env
# Edit dev.env, then:
make use && make dev

  10. Reset MindsHub installation with `make flush`

    main

    Use make flush to perform a clean reset of your local environment. This is useful for testing installation flows or recovering from a broken installation.

    Warning: This command deletes your conversations and saved keys. It removes:

    • The local runtime (cowork-server uv tool and backend/*/.venv).
    • App state in ~/.anton (provider keys).
    • App state in ~/.cowork (database, hermes, projects).

    To skip the confirmation prompt, pass FORCE=1:

    make flush FORCE=1

    After flushing, run make setup or launch the app to reinstall everything.

  11. Reset MindsHub environment with make flush

    main

    The make flush command is used to perform a clean installation or recover from a corrupted installation.

    Warning: This operation will delete your conversation history, saved keys, and application state. It removes the local cowork-server (uv tool) and backend/*/.venv, and deletes ~/.anton (provider keys) and ~/.cowork (database, hermes, projects).

    To skip the confirmation prompt, use the FORCE=1 environment variable.

    make flush FORCE=1

  12. Reset MindsHub local environment

    main

    If you need to start from scratch or recover from a broken installation, use the make flush command.

    Warning: This will delete your local runtime (cowork-server and backend/*/.venv) and erase all app state in ~/.anton (provider keys) and ~/.cowork (database, Hermes, and projects). This action deletes all your conversations and saved keys.

    To skip the confirmation prompt, use:

    make flush FORCE=1