#3232nd of 49 in Desktop AI appsFor data and documents, data analysis, science, health and research, research and papers
open-science
Local-first research workbench with provenance tracking and portable .science exports
- Stars
- 5.5k
- License
- Apache-2.0
- Last commit
- Nov 2026
- Last release
- Nov 2026
- Language
- TypeScript
2 on 30 Jun 20265,515 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
AIPOCH Open-Science is a desktop app (macOS, Windows, Linux) with a standalone CLI and Web UI, also installable via npm on Node.js 22.13.0 or later. An agent works over papers and project data, with Python/R notebooks, Zotero import, literature screening and a Provenance view of inputs, code and execution history. Completed sessions export as portable read-only .science packages that others can import and replay.
Who it is for: Researchers wanting reproducible, shareable AI analysis on their own machine
Strengths
- Exports sessions as .science packages; imports are read-only and run no code or credentials
- Provenance view shows inputs, code and execution history per figure or table version
- Installers for macOS, Windows, Linux (x64/ARM64), plus Homebrew, npm and standalone CLI
- Benchmark results published with caveats that they are local runs
Weaknesses
- Benchmark scores are AIPOCH's own local runs, not official leaderboard submissions
- Rerunning an imported result needs a complete recipe, inputs and an available runtime
- Remote Web sessions cannot read or write arbitrary local package paths
- Research narration requires a separately installed local voice model
What it needs
- Needs Node.js 22.13.0+ (npm install only), Python or R (optional, for notebooks)
- Models: Codex subscriptions, Claude, DeepSeek, Gemini, configurable provider via API key and endpoint
Quick answers
Is open-science open source?
Yes. Its code is on GitHub under the Apache-2.0 license.
Also in Desktop AI apps
See all 49| Rank | Project | Score |
|---|---|---|
| 1 | cherry-studioDesktop client for chatting with many LLM providers, with agents and knowledge bases | 92 out of 100 |
| 2 | dsh-desktopNative Windows and macOS desktop client for DeepSeek Harness, built on plugins | 90 out of 100 |
| 3 | screenpipeRecords screen and audio locally, makes it searchable for you and your agents | 90 out of 100 |
| #4 | prompt-optimizerPrompt rewriting and testing tool for web, desktop, Chrome and MCP | 86 out of 100 |
| #5 | aionuiDesktop and WebUI app that runs a built-in agent plus external CLI agents | 86 out of 100 |