# AIPOCH Open-Science

> AI research workbench for scientists — Local-first AI research workbench: agents run Python/R, every result keeps provenance

- Page: https://tesign.com/en/item/aipoch-open-science/
- JSON: https://tesign.com/en/item/aipoch-open-science/index.json
- Korean Markdown: https://tesign.com/item/aipoch-open-science/index.md
- Generated: 2026-09-17 01:56 UTC

## RANKS

- SCIENCE #2 (As of 2026-09-16 21:00 UTC)

## Numbers

- 4,196 stars — checked on GitHub 2026-09-15 19:35 UTC (+3 observed since the check)
- 7-day +26 observed via GH Archive (as of 2026-09-16 21:00 UTC)
- 24h +3 ★ · 30d +186 ★ (as of 2026-09-16 21:00 UTC)
- Star total = the value last checked on GitHub (stars_checked_at) + increases observed via GH Archive since. The 24h · 7d · 30d gains are GH Archive hourly events summed to the reference time (as_of). No score of ours.

## AT A GLANCE

- LICENSE: Apache-2.0 (permissive) — https://spdx.org/licenses/Apache-2.0.html
- USAGE: Commercial use, redistribution OK. Keep notices; mark changes.
- OPEN SOURCE: YES
- LANGUAGE: TypeScript
- PLATFORM: macos · windows · linux · cli
- CATEGORY: SCIENCE · AI
- Tags: ai-research · provenance · python · r · electron · mcp
- How to start: Install to use
- SOURCES: GitHub https://github.com/aipoch/open-science
- INSTALL: https://aipoch.com/open-science

## ACTIVITY

- Last commit: 2026-09-15 15:52 UTC
- Latest release: v0.30.0 (2026-09-15)
- Contributors: 17
- Open issues (incl. PRs): 43
- Made by: an organization
- Checked on GitHub: 2026-09-15 19:35 UTC

## Signals and evidence

- BACK CATALOG

## TESIGN TAKE

Its focus is recording where each result came from, with versions and replay verification — while model costs stay separate and prompts plus attached context still go to whichever provider you pick.

## WHY IT MATTERS

It gathers research AI work that usually scatters across chat windows, notebooks, scripts and reporting tools into one desktop app. A five-step first run installs Claude Code, OpenCode, Codex or CodeBuddy as the agent runtime (no Node.js or admin password needed) and connects a model via a built-in cloud provider, a custom gateway or a Claude/Codex subscription login. It includes Python and R kernels, literature search across Europe PMC, OpenAlex and arXiv, 24 built-in data connectors and 22 featured skills, and every generated table or figure is kept as an immutable version with checksums, execution history and inputs that can be replayed in an isolated environment for verification. An Electron app for macOS, Windows and Linux; v0.29.0 (2026-09-13), Apache-2.0.

## BUILD FROM THIS

- The headless CLI and Node.js SDK share the desktop's local daemon, projects and permissions, so you could build an unattended pipeline that re-analyses a fixed dataset on a schedule and emits a report with verification records. Custom MCP connectors can bring in an in-house database.

## WHO IT'S FOR

- Computational researchers in bioinformatics, statistics, chemistry — agents run code, search literature, produce tables and figures
- Labs held to reproducibility standards — artifact versions, checksums and replay verification records
- Teams on Slurm clusters — notebook runs submitted over SSH

## START IN 5 MINUTES

```
# brew install --cask open-science   # macOS
# # Windows x64 installer and Linux x64 AppImage/.deb are under Assets on the latest GitHub release
# # First run, five steps: Environment → Data location → Agent runtime (Claude Code/OpenCode/Codex/CodeBuddy) → Model provider connection test → Notebook runtime (optional)
# # New project → describe goal, input data and how to check the result in a session → run, then open the Provenance view
```

## CAVEATS

- Apache-2.0 — the software is free with no seat licence (homepage), and the homepage states it is not affiliated with Anthropic
- Model usage is billed separately: built-in cloud providers and custom gateways need your own API key, or reuse a Claude/Codex subscription login; prompts and attached context are sent to the provider you choose
- Electron desktop app, v0.29.0 (2026-09-13) — the README says provider lists, catalogs and UI change between versions; the #1 BiomniBench-DA Public 50 claim is the project's own report

## RECEIPT

- FIRST SEEN BY TESIGN: 2026-09-13 09:38 UTC
- AT SOURCE: 2026-07-03 02:50 UTC
- KEPT: 2026-09-14 19:27 UTC
- Published on TESIGN: 2026-09-14 19:27 UTC
- Text last updated: 2026-09-15 04:36 UTC
- ◌ BACK CATALOG: Reconstructed from archive data, not a live discovery.

## How to read this

- This file was produced by the same build, from the same data, as the tesign.com item page. The text is editorial; the numbers are stored observations.
- null in the JSON means not observed — never zero. The Markdown writes [unconfirmed] for it.
- Star total = the value last checked on GitHub (stars_checked_at) + increases observed via GH Archive since. The 24h · 7d · 30d gains are GH Archive hourly events summed to the reference time (as_of). No score of ours.
- Reconstructed from archive data, not a live discovery.
- A summary, not legal advice.

[Image] https://tesign.com/img/aipoch-open-science-3bdef77a47.jpg