AIPOCH Open-Science v0.34.1

AIPOCH Open-Science v0.34.1

A connector-and-reliability release: CELLxGENE Discover and Alliance of Genome Resources arrive as new connectors, the variants and omics connectors expand with MaveDB and Metabolomics Workbench coverage, the workspace library preview gains an Inbox tab, and notebook reliability improves across cells and sessions.

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.

v0.34.1 grows the connector catalog and tightens everyday reliability. A new CELLxGENE Discover connector searches public single-cell datasets and collections with rich metadata filters, published-version inspection, and download links, plus CellGuide cell-type descriptions. A new Alliance of Genome Resources connector brings cross-species model-organism gene knowledge — orthologs, disease models, phenotypes, alleles, and expression across eight species. The variants connector gains MaveDB functional scores, and the omics connector deepens Metabolomics Workbench coverage. The model picker adds gpt-6.1-sol. In the workspace, the library preview gains an Inbox tab for the pending queue with batch acceptance, and the empty state offers a .science import entry. Notebook fixes preserve interpreter state across cells, share sandbox tools across sessions, and explain loopback gateway bind failures. Platform fixes restore Linux KWallet startup and allow Windows reinstalls after removing an old parent installation.

✨ Highlights

• New connector pair. A CELLxGENE Discover connector discovers public single-cell datasets, and an Alliance of Genome Resources connector covers cross-species model-organism genes. (#3178, #3136)
• Variant and omics expansion. MaveDB functional scores join the variants connector, and the omics connector expands its Metabolomics Workbench coverage. (#3131, #3139)
• Library preview Inbox. The workspace library preview gains an Inbox tab for the pending queue, with accept, dismiss, undo, and batch acceptance. (#3174)
• gpt-6.1-sol. The model picker adds gpt-6.1-sol without changing existing defaults. (#3157)

🚀 New Features

• CELLxGENE Discover connector: search public single-cell datasets and collections, filter by organism, tissue, disease, assay, or cell type, inspect published versions, and get file formats, sizes, and download URLs, plus CellGuide cell-type descriptions and marker genes. (#3178)
• Alliance of Genome Resources connector: search and summarize genes across human, mouse, rat, fly, worm, zebrafish, yeast, and frog — orthologs, disease models, phenotype annotations, alleles, expression, and disease-term associations. (#3136)
• MaveDB functional scores in the variants connector: score-set search and metadata, assay-specific functional scores, VRS variant mappings, and experiments. (#3131)
• Expanded Metabolomics Workbench coverage in the omics connector: study records, samples, experimental factors, analysis metadata, and compound structures and cross-references. (#3139)
• gpt-6.1-sol as a model option, with defaults unchanged. (#3157)
• Inbox tab in the workspace library preview: browse the pending queue with search and pagination, accept or dismiss individual items with undo, and accept batches by explicit page selection. (#3174)
• A .science import entry in the workspace empty state for importing research packages. (#3170)
• Session packages can export acknowledged sensitive content with explicit confirmation. (#3164)
• Local diagnostics exports preserve troubleshooting evidence for support. (#3143)
• Feedback send shortcuts in session plans. (#3125)
• Host status shown in the composer compute menu. (#3124)

🔧 Improvements

• Run previews link directly to the relevant messages. (#3120)
• Skill drafts are protected and skill interactions clarified. (#3142)
• Journal mapping hints are simplified and journal import column roles clarified. (#3172, #3169)

🐛 Bug Fixes

• Notebook and runtimes — interpreter state is preserved across cells (#3129); sandbox npm global tools are shared across sessions (#3160); loopback gateway bind failures are explained in plain language (#3138); Python and R scientific lineage capture is hardened (#3163).
• PDF and preview — native figure and table content is preserved during structure extraction (#3162); PDF preview workers run in browser clients (#3135); reference URLs appear in the detail view (#3151); annotation selection clears before the reader closes (#3156); merged-table preview styling and the pinned-column overflow shadow are corrected (#3165, #3167).
• Sessions and packages — follow-up permissions survive agent restarts (#3147); boolean metadata is preserved during package export (#3133).
• Journals — journal filters are validated and alias columns recognized (#3145); invalidated journal entry reads are retried (#3148).
• Platform — Windows allows reinstall after removing an old parent installation (#3176); Linux KWallet startup is restored (#3168); credential identity accepts completed SQLite journals (#3126).

📦 Install

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 can update itself in place.

Download the appropriate package from the Assets section below, or from the official download page:

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

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 v0.34.0, plus: the CELLxGENE Discover and Alliance of Genome Resources connectors; MaveDB functional scores in the variants connector; expanded Metabolomics Workbench coverage; the gpt-6.1-sol model option; the library preview Inbox; and the .science empty-state import entry.
• 🚧 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 and RO-Crate packaging, while full-session replay and solver-exact equivalence remain open; conditional package restore is an exported script workflow; and 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

• Verification checks a captured version, not a whole session. Full-fidelity session replay and solver-exact equivalence guarantees remain open, and starting a check from the web client currently requires the desktop app.
• Sensitive package evidence is opt-in. Session diagnostics include sensitive package evidence only after explicit confirmation, and the export stays local until you share it.
• 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.
• 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.
• 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.
• 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, @wen2zhou, @nasus2002, @Ada.Liu, and everyone in Discord, X, and Discussions.

---

Full Changelog: https://github.com/aipoch/open-science/commits/v0.34.1