AIPOCH Open-Science v1.0.5

AIPOCH Open-Science v1.0.5

Connectors reach native scientific APIs and official file discovery, the literature library tracks reading status, session packages export and import in the browser, local service lifetime follows explicit holds, updates gain manual entry points and an installer reveal, and network permissions, OpenCode 2.x compatibility, and notebook startup paths are fixed across platforms.

AIPOCH Open-Science is an open-source, local-first, model-agnostic AI research workbench for scientists and researchers. It enables reproducible, inspectable research across models with scientific AI agents, Python and R execution, scientific data connectors, and cross-platform support for macOS, Windows, and Linux.

v1.0.5 follows v1.0.4 with a breadth-and-polish release. The scientific connectors extend beyond unified queries to native source APIs: Europe PMC gains direct Boolean, core and abstract search with explicit pagination, alongside HGNC gene metadata, ChEA3 enrichment, bioDBnet conversion and orthology, Cellosaurus RDF knowledge-graph queries and STR similarity, and official file discovery for NHANES, PhysioNet, openFDA and CTD — all opt-in, returning official URLs and provenance. The literature library grows a manually managed reading status with a dedicated column, inline and bulk editing, and filtering. Local Web deployments can export and import .science session packages entirely in the browser. The local service now runs only while an app connection or explicit start hold keeps it alive, so closing one window no longer disturbs unrelated background work. Manual update entry points, an installer reveal, and a security policy link round out the surfaces, with fixes across network permissions, OpenCode 2.x, and notebook startup on macOS and Windows.

✨ Highlights

• Scientific connectors reach native source APIs. Europe PMC gains direct Boolean, core and abstract search with explicit pagination, alongside new HGNC gene metadata, ChEA3 enrichment ranks, bioDBnet conversions and orthology, Cellosaurus RDF knowledge-graph queries, STR similarity, and official file discovery for NHANES, PhysioNet, openFDA and CTD — all opt-in and returning official URLs with provenance rather than opaque downloads. Queries and identifiers you submit — search terms, gene names, DOIs — are sent to the respective public services, and access-gated sources are listed with official access instructions only; the app never asserts your authorization. See the scientific-data API documentation for scope and access notes. (#3869)
• Track reading progress in the literature library. References can be marked Unread, In progress, or Finished reading; the status shows in a dedicated column, edits inline or from the details page, updates in bulk across pages and exclusions, and filters the library with consistent counts and selection scope. (#3897, #3907)
• Session packages work in the browser. Local Web deployments can export a .science package as a download and import one through an upload picker — no CLI fallback needed when working entirely in the browser. An exported package contains the conversations, files, notebook records, and provenance you selected, and browser transfers run between your browser and your own local service. See the security guidance on treating imported packages as untrusted content until reviewed. (#3911)
• Service lifetime follows your intent. Keeping the app open or running open-science start holds the local service alive; closing one window no longer stops it or unrelated background work. Plain stop releases your hold, and stop --force requests a whole-service shutdown. While a hold is active, the service and its background work keep running until you release the hold or force the shutdown. See the CLI service lifecycle documentation for the hold and shutdown semantics. (#3769)

🚀 New Features

• The Home page gains a use case gallery: browse published examples, open their details, and import a verified session package with download progress and cancellation. (#3901)
• Manual update entry points: persistent download and instruction links in Settings and update surfaces, plus a read-only command that reveals the confirmed installer. (#3902)
• Settings → About links the project's security policy. (#3906)
• Authorized directories render as a compact folder list with read/write badges and grant-time tooltips, instead of raw JSON. (#3871)

🔧 Improvements

• DOCX reports preserve intended table styles, borders and page layout during generation and preview, and a maintained, template-based Markdown-to-DOCX conversion workflow is available on demand. (#3811)
• Publication-check results present one short action with Crossref attribution and last-check time. (#3896)
• Reference details consolidate tags and type into the header, with repetitive explanations removed. (#3910)
• Imported session package and replay surfaces drop redundant diagnostics and give their actions consistent, visible boundaries. (#3908)
• Windows managed R requests for Bioconductor packages now explain the explicit BiocManager route instead of failing with a generic Conda error. (#3918)

🐛 Bug Fixes

• Full-access sessions admit public network destinations without approval prompts, and downgrading Full closes existing connections. (#3898)
• Settings → Network immediately reflects global domain grants made from approval dialogs. (#3900)
• OpenCode 2.x runtimes start and work again, with the 1.x path retained. (#3903)
• macOS notebooks find CLI tools whose Node runtime lives outside the app's inherited PATH. (#3916)
• Windows chats no longer wait on protection configuration before a provider request starts. (#3922)
• Notebooks read application-managed skill reference files without a folder-access prompt. (#3920)
• Collapsing tool groups preserves your reading position. (#3873)
• Wide Markdown tables reveal their scrollbar on hover and keyboard focus. (#3875)

📦 Install

Desktop client. Requirements: macOS 12+ (Apple Silicon or Intel), Linux x64 or ARM64, or Windows 10/11 x64. On first run, the onboarding wizard checks the environment and can install and configure an app-managed agent runtime. Once installed, the app updates itself in place.

Download the appropriate package from the Assets section below, or from the official download page. On macOS, the desktop app is also installable via Homebrew: brew install --cask open-science.

Platform	Package
macOS (Apple Silicon)	DMG for ARM64
macOS (Intel)	DMG for x64
Linux	AppImage or Debian package for x64 / ARM64
Windows	Code-signed installer for x64

Command line via Homebrew (macOS and Linux). The standalone CLI and local Web UI live in the official tap — bundled Node runtime, no separate Node.js needed, and upgrades through the normal brew flow:

brew install aipoch/tap/open-science
open-science start --no-open
open-science url

Command line via npm. With Node.js 22.13+, install the CLI (and Task SDK) globally; updates then follow your normal npm upgrade flow:

npm install --global @aipoch/open-science
open-science --help

Command line standalone (all platforms, no toolchain). The Assets below also include standalone CLI archives — macOS and Linux arm64/x64 tarballs and a Windows x64 zip — with a bundled Node runtime, so they install and run without Homebrew, npm, or a local Node.js. Linux packages additionally arrive as standalone .deb and .rpm installers.

macOS — first launch. Official release builds are Developer ID signed and notarized by Apple, so they open like other trusted applications. A locally built copy is notarized differently and may require approval through macOS Privacy & Security.

Windows — first launch. Official release builds are code-signed; Windows recognizes the publisher and no SmartScreen warning appears. Verify that the package came from the official release page before continuing.

Build from source instead:

npm install
npm run build:mac   # or: build:linux / build:win

🧭 What's in this release (maturity)

• ✅ Implemented: everything shipping in v1.0.4, plus native scientific source APIs and official file discovery in the connectors, literature reading-status management and filtering, browser session-package export and import on local Web deployments, explicit-hold service lifetime, manual update entry points with installer reveal, the security policy link, and the fixes described above.
• 🚧 Partial: provider choice remains constrained by the active framework's endpoint compatibility; remote compute covers direct SSH and Slurm (cloud-GPU submission is not built yet); WSL2 Bash remains an opt-in Windows x64 preview; the skills commons ships marketplace discovery but not cross-machine forking or user-facing version pinning; reproducibility ships replayable per-version verification, session replay, and RO-Crate packaging, while solver-exact equivalence remains open; review is opt-in and record-scoped.
• 🗺️ Roadmap: a unified model gateway, cross-machine skill forking and version pinning, cloud-GPU execution, stronger sandboxing and credential isolation, and collaborative research workflows.

🐢 Known Limitations

• Replay reproduces recorded steps, it does not re-execute them. Solver-exact equivalence guarantees remain open; replayable verification remains the deterministic check, and starting a check from the web client currently requires the desktop app.
• Imported .science sessions are read-only. They stay inspectable and referenceable, and forking them into a writable continuation remains available; execution and continuation remain disabled on the import itself.
• PDF annotations are bound to the file version they annotate. Notes follow the file across projects and sessions where that version is shared, but they do not sync across machines.
• Bookmarks belong to a single session. They do not follow branch switches, join .science package exports, or sync across machines — forking copies them into the new session with fresh identities.
• WSL2 Bash is a preview. It is opt-in on Windows x64 only; PowerShell remains the default shell runtime.
• Remote compute covers direct SSH and Slurm. Cloud-GPU submission is not built yet.
• Editable artifacts cover text formats. Binary formats stay read-only, and editing always publishes new versions.
• Network sandboxing covers the app's notebook and compute runtimes, not the whole system. On Windows, the boundary applies only after the sandbox's one-time administrator setup.
• Provider choice is per framework, not one unified gateway. The available protocol depends on the selected agent backend, and remote endpoints must use HTTPS.
• Code reconstruction is LLM-generated. It does not replace deterministic reproduction; replayable verification is the deterministic check.
• The reviewer is opt-in and record-scoped. It does not replace domain-specific validation of citations, units, statistics, or methods.
• No local GPU compute backend.
• No multi-user real-time collaboration.

🙏 Acknowledgements

Thanks to @ewen-poch, @daanveer-tech, @nasus2002, @roxi3906, @justemu, @wen2zhou, @yshtonytc, @LJLJ111 and everyone in Discord, X, and Discussions.

---

Full Changelog: https://github.com/aipoch/open-science/commits/v1.0.5