Repolode
Submit
aipoch

aipoch/open-science

Local-first, model-agnostic desktop AI research workbench for scientists, with agents, Python/R notebooks, scientific connectors and traceable artifacts.

View on GitHubWebsite
Stars
5,249
Forks
522
Open issues
57
Last commit
19 hours ago

What is it?

AIPOCH Open-Science is a desktop research workbench for scientists that runs on macOS, Windows and Linux. Users create a project, describe a research goal in plain language, and agents can read attached files, search the web, run Python or R code in a shared Notebook, query scientific data sources, and produce reports, tables and figures.

The app is local-first: project data, settings, artifact versions and provenance records are stored on the local machine. It works with cloud model providers, custom OpenAI-compatible gateways including local servers such as Ollama and vLLM, and Claude or Codex subscription logins. Agent backends can be Claude Code, OpenCode, Codex or CodeBuddy.

Each generated artifact is stored as a checksummed version with its producing code, inputs and execution history where available. Completed sessions can be exported as portable `.science` packages for review or handoff, and imported records are read-only. The project also offers a headless CLI and Node.js SDK that share the same local service.

Key features

  • 01Project-based research sessions where agents read files, search the web and run code.
  • 02Python and R Notebook execution in local environments or over SSH and Slurm.
  • 0323 built-in skills and 36 built-in scientific connectors, with custom MCP connector support.
  • 04Literature library with PDF reading, open-access full-text lookup and BibTeX/RIS export.
  • 05Immutable artifact versions with provenance views and RO-Crate 1.1 export.
  • 06Portable `.science` packages for exporting and importing completed research sessions.

Use cases

  • Reproducing a published data analysis and comparing regenerated figures and tables with the paper.
  • Running literature reviews with search, PDF reading and citation export in one workspace.
  • Handing off or archiving a completed computational study with its code, data versions and execution records.
  • Running Python or R analysis on local data while keeping sensitive files on the local machine.

Quick start

  1. 1

    Download the app from the official download page

  2. 2

    Install on macOS with Homebrew

    brew install --cask open-science
  3. 3

    Complete the first-time setup wizard and test your model connection

Things to know

  • Requires a configured model provider connection, such as an API key, to run agents.
  • The README does not state whether it runs offline with local models.

Topics

#ai-agent#ai-for-science#research-tools#reproducible-research#local-first#mcp#jupyter-notebook#desktop-app#agent-skills#ai-research

FAQ

What is open-science?
Local-first, model-agnostic desktop AI research workbench for scientists, with agents, Python/R notebooks, scientific connectors and traceable artifacts. AIPOCH Open-Science is a desktop research workbench for scientists that runs on macOS, Windows and Linux. Users create a project, describe a research goal in plain language, and agents can read attached files, search the web, run Python or R code in a shared Notebook, query scientific data sources, and produce reports, tables and figures.
Is open-science free to use? What license does it have?
open-science is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is open-science still maintained?
Actively maintained. The last commit to open-science was on Oct 19, 2026.
What can I use open-science for?
Reproducing a published data analysis and comparing regenerated figures and tables with the paper. Running literature reviews with search, PDF reading and citation export in one workspace. Handing off or archiving a completed computational study with its code, data versions and execution records. Running Python or R analysis on local data while keeping sensitive files on the local machine.
How do I install open-science?
Download the app from the official download page Install on macOS with Homebrew: brew install --cask open-science Complete the first-time setup wizard and test your model connection
Where can I get open-science?
The source code and installation instructions are on GitHub: https://github.com/aipoch/open-science

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects