Repository navigation
v0.4.1
Open Science v0.4.1
This release is about reaching your workbench from anywhere on your machine, and owning the agent backend underneath it. You can now open the same workbench in a browser (an optional localhost web UI), run it headless with no desktop window, and drive that backend as a background service from an
open-sciencecommand line. Settings also gains full control over the app-managed agent runtimes — install, switch, and uninstall Claude Code or OpenCode — plus two new model providers, smarter composer mentions, and a round of reliability hardening. All self-hosted, on your own machine.
Open Science is an open-source, model-agnostic AI workbench for scientific discovery — a self-hosted desktop app pairing a plan-and-execute agent with a persistent, sandboxed compute kernel and durable project/session storage.
This is a reach-and-control release: it makes the workbench accessible from a browser and the terminal, hands you install/uninstall control over the agent-framework runtimes shipped in v0.4.0, and hardens the OpenCode, data-root, and conversation-recovery paths. The public skills commons and a fully model-agnostic gateway remain ahead on the Roadmap.
✨ Highlights
- Open the workbench in a browser. The desktop backend can optionally serve the same renderer to a browser on your own computer, bound only to
127.0.0.1behind an authenticated URL, and can run headless — backend, tray, agent runtime, and notebook kernels without opening an Electron window. Off by default; start it withnpm run dev:web/npm run dev:headlessin development, or from the CLI on an installed build. (#205, #207) - Drive it from the command line. A bundled
open-scienceCLI manages the headless backend as a background daemon —start,status,url,stop— so the browser is the only UI and closing your terminal doesn't stop active sessions. From a packaged app, Settings → General → Command line tool → Install command puts anopen-sciencelauncher on your PATH that runs the bundled CLI on the app's own runtime, with no separate Node.js install. (#206, #211) - Manage the agent backend yourself. Settings cards now let you install, switch, and uninstall the app-managed Claude Code and OpenCode runtimes. Uninstall removes only a runtime the app downloaded and manages itself (a PATH/npm/homebrew binary is never touched), and when you remove the active backend the app auto-switches to the other one if its runtime is version-verified. The uninstall action is always visible, with an inline explainer for when it doesn't apply. (#193, #204)
- Two new model providers. Add OpenRouter — a multi-vendor aggregation gateway reaching top models across Anthropic, OpenAI, Google, and others behind one key — and Xiaomi MIMO, both over Anthropic- and OpenAI-compatible endpoints. (#198)
🚀 New Features
- Localhost web UI + headless backend mode — serve the renderer to a loopback-bound browser behind an authenticated URL, and run the backend/tray/agent/notebook without an Electron window; a
OPEN_SCIENCE_WEB_PORTselects the port and explicit quit still shuts down agent and notebook processes normally. (#205, #207) open-scienceCLI daemon + PATH launcher — manage the headless backend as a background service (start/status/url/stop, with--no-open,--port,--json,--app-path,--config-root), and install the launcher to your PATH directly from the packaged app so it runs on the app's own runtime. (#206, #211)- Manage agent-framework runtimes from Settings — select the active backend and install or uninstall the app-managed Claude Code / OpenCode runtime from the status cards, with a confirmation dialog, guarded to managed installs, and an active-backend auto-switch to a version-verified alternative. (#193, #204)
- OpenRouter and Xiaomi MIMO providers — two new official vendors; OpenRouter ships a curated cross-vendor catalog and sits last in the picker as an aggregation gateway. (#198)
- Fuzzy
/and@composer mentions — the skill (/) and file (@) mention popups now fuzzy-match and rank results with match highlighting, instead of requiring a prefix. (#200) - Resilient nested-bundle skill import — import a zip of zips, skip individual bad skills instead of failing the whole batch, and import multiple packages in one pass, with a per-file size cap. (#201)
🔧 Improvements
- Conversations that outgrow the model's request-size limit now recover instead of getting stuck — an oversized turn with unstrippable media is caught and compacted rather than failing (
media_unstrippable). (#199) - OpenCode hardening: config and auth isolation, session cleanup, and install are tightened, and a baseline OpenCode is installed on non-AVX2 x64 hosts with the app's home config isolated from your own. (#192, #195)
- Provider-not-found errors from the agent are now surfaced as an actionable prompt message instead of an opaque failure. (#194)
- In-place updates now stop the backend process tree before installing, so an update can't collide with a still-running backend. (#196)
- The first-run data-root notice is softened from a red error to an amber advisory to match its severity. (#202)
🐛 Bug Fixes
- The false "Data folder not found" nag is fixed for non-ENOENT stat errors — notably non-ASCII (e.g. CJK) Windows install paths, which previously reported a missing data folder that was actually present. (#203)
📦 Install
Requirements: macOS 12+ (Apple Silicon or Intel), Linux x64, or Windows 10/11 x64. On first run the app's onboarding wizard checks your environment and can install and configure a Claude-compatible model backend for you. Once installed, the app updates itself in place.
Download the build for your platform from the Assets below:
| Platform | File |
|---|---|
| macOS (Apple Silicon) | open-science-0.4.1-mac-arm64.dmg |
| macOS (Intel) | open-science-0.4.1-mac-x64.dmg |
| Linux | open-science-0.4.1-linux-x86_64.AppImage / .deb |
| Windows | open-science-0.4.1-win-x64-setup.exe |
macOS — first launch. Release builds are Developer ID signed and notarized by Apple, so they open like any other app. (A .app you build yourself is not notarized — see README → Building From Source.)
Windows — first launch (unsigned build). No Authenticode certificate yet, so SmartScreen shows a bypassable "unrecognized app" prompt (More info → Run anyway).
Build from source instead:
npm install
npm run build:mac # or: build:linux / build:win
🧭 What's in this release (maturity)
Aligned with the PRD and Roadmap:
- ✅ Implemented: desktop shell (Electron + React + TS) with a tray/single-instance lifecycle, now also reachable as an optional localhost web UI, a headless backend mode, and an
open-scienceCLI daemon (installable to PATH from the packaged app); ACP agent runtime on a pluggable framework backend (Claude Code default, OpenCode alternative) whose app-managed runtimes are installable, switchable, and uninstallable from Settings; an opt-in reviewer with a bounded fix loop; one persistent Python kernel with replayable run history; project + session persistence; artifacts MCP server; file renderers + notebook/session-notebook viewers + offline molecule viewer; multi-provider config (built-in vendors Claude, DeepSeek, Zhipu AI (GLM), MiniMax, Kimi (Moonshot), Kimi For Coding, Xiaomi MIMO, plus the OpenRouter aggregation gateway, custom gateways, and local Claude) over Anthropic-compatible endpoints on any backend and OpenAI-compatible endpoints via OpenCode; signed-and-notarized macOS packaging plus Windows/Linux, with in-place auto-update and a configurable data-storage location; file-based agent skills with a fuzzy/selector and resilient batch import; 24 built-in connectors (23 upstream-aligned life-science connectors at 200+ tools + the molecule viewer) plus custom MCP servers; a guided first-run onboarding wizard; and per-conversation approval profiles. - 🚧 Partial: the new CLI/browser surface is a first entry point, not yet a scriptable/embeddable SDK; provider choice is still split across two backends (Claude Code speaks Anthropic; OpenCode adds OpenAI-compatible) and gated on the selected framework's endpoint compatibility rather than one unified gateway; skills are local and file-based (no shared/public commons, cross-machine forking, or version pinning yet); the reviewer verifies a turn's own record but does not yet apply domain-specific citation/units/statistics rubrics.
- 🗺️ Roadmap (not in this build): public skills commons and savable specialist roles (rest of Phase 3); a fully model-agnostic gateway with unified per-agent routing and a full CLI/SDK entry point (rest of Phase 1); artifact versioning + provenance chain, additional kernels (R), and specialist sub-agents (Phase 2); remote compute and the rest of the security stack — network sandbox, credential vault, scoped permission tiers (Phase 4).
🐢 Known Limitations
- The web UI is loopback-only. It binds to
127.0.0.1behind an authenticated URL by design; there is no LAN/remote-access mode in this build. Exposing the workbench beyond localhost is separate, deliberately-scoped work. - The CLI is an entry point, not an SDK. It manages the headless backend as a daemon; a scriptable/embeddable programmatic interface is still open Phase 1 work.
- Provider choice is per backend, not one unified gateway. Claude Code reaches Anthropic-compatible endpoints; OpenCode adds OpenAI-compatible ones (OpenRouter and Xiaomi MIMO speak both). The fully model-agnostic gateway with unified per-agent routing is still open Phase 1 work.
- Switching backends can't carry a live session over. Open conversations keep their messages and get replayed to the new backend, but in-flight tool state is not migrated — switching starts a fresh agent session.
- Skills are local only. You can create, edit, and import them on one machine, but there is no shared/public commons, no cross-machine forking, and no user-facing version pinning yet.
- The reviewer is opt-in and record-scoped. It traces a turn's own claims against its transcript/logs/artifacts; it is off by default and does not yet run domain-specific citation/units/statistics checks.
- Windows builds are unsigned. Windows has no Authenticode cert yet, so it shows a bypassable SmartScreen prompt on first launch. (macOS is notarized.)
- No provenance chain / artifact versioning yet. Lands in Phase 2.
- No local compute backend. The bundled biomodel skills (AlphaFold2, ESMFold2, DiffDock, …) ship as reference/instruction docs; this build can't execute their GPU/model tooling locally.
- No multi-user / real-time collaboration. Sharing is via export/import.
🙏 Acknowledgements
Thanks to @ewen-poch, @wen2zhou, and @imjszhang (first contribution — the localhost web UI and CLI daemon), and to everyone in Discord, X, and Discussions.
What's Changed
- fix: harden OpenCode config isolation, session cleanup, and install by @ewen-poch in #192
- fix(acp): surface provider not-found errors with an actionable prompt message by @ewen-poch in #194
- fix: install baseline OpenCode on non-AVX2 x64 and isolate home config by @ewen-poch in #195
- feat(settings): manage agent runtimes — select, install, and uninstall from cards by @ewen-poch in #193
- fix(update): stop backend process trees before in-place install by @ewen-poch in #196
- feat(providers): add OpenRouter and Xiaomi MIMO official vendors by @ewen-poch in #198
- feat(composer): fuzzy matching for / and @ mention popups by @ewen-poch in #200
- feat(skills): make nested-bundle skill import resilient by @ewen-poch in #201
- fix: recover conversations that outgrow the request size limit (media_unstrippable) by @ewen-poch in #199
- fix(onboarding): soften data-root notice from red to amber advisory by @ewen-poch in #202
- fix(storage): stop false "data folder not found" nag on non-ENOENT stat errors by @ewen-poch in #203
- feat(settings): always show agent-card uninstall with an inline explainer by @ewen-poch in #204
- feat(web): add localhost web UI and headless backend mode by @imjszhang in #205
- fix(web): review follow-ups for localhost web UI (#205) by @ewen-poch in #207
- feat(cli): manage the headless backend as a daemon by @imjszhang in #206
- feat(cli): install a terminal launcher for the packaged app and fix daemon follow-ups by @ewen-poch in #211
- chore: bump version to 0.4.1 by @ewen-poch in #213
New Contributors
- @imjszhang made their first contribution in #205
Full Changelog: v0.4.0...v0.4.1