v0.8.0

Open Science v0.8.0

Branch a research conversation without losing the original path, and trace every generated result through immutable artifact versions and inspectable provenance.

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 persistent compute, durable project/session storage, and inspectable research artifacts.

v0.8.0 adds two connected foundations for exploratory research. Editing an earlier prompt now creates a selectable message branch while preserving the original conversation, and generated outputs receive immutable, checksummed versions linked to their producing branch and available code, execution, input, environment, and review evidence.

✨ Highlights

• Immutable artifact versions and provenance. Each generated output can now be inspected as a specific version with checksummed content and linked code, execution, input, environment, branch, and review evidence. (#481)
• Message branches for alternative research paths. Edit and resend a completed prompt to explore a different direction without deleting the original downstream conversation; revision controls switch between the alternatives, and branch state persists across restarts. (#481)
• Model and reasoning effort in one picker. The composer now keeps model choice and the selected model's supported reasoning levels together, with live synchronization across supported agent frameworks. (#468)
• Preview navigation that stays oriented. The active preview tab remains visible through panel resizing, while long names preserve their identifying suffix and extension across previews, mentions, attachments, and artifact suggestions. (#478)
• A refreshed cross-platform identity. New light and dark application artwork reaches packaged apps, windows, the Dock, and system trays, with Windows-specific rendering tuned for native sizes and display scaling. (#479, #484)
• More reliable remote results with relocated storage. Harvested SSH compute results now follow the configured data-storage location, so jobs remain attachable after moving large project data away from the default location. (#480)

🚀 New Features

• Durable provenance records — generated artifacts gain session-scoped lineages, immutable versions, checksummed content, producer and input evidence, environment inventories, producing-branch context, version-scoped review evidence, crash recovery, and migration-aware validation. (#481)
• Persistent message branching — editing a completed user message forks the conversation from that point, keeps the original later turns available, and preserves branch-specific activity, attachments, files, and artifact ownership. (#481)
• Combined composer model control — switch the model and its reasoning effort from one compact menu; available levels adapt to the selected model's capability profile, and changes apply immediately where the active framework supports live updates. (#468)
• V2 application identity — light and dark brand assets now cover packaged and runtime surfaces on macOS, Windows, and Linux, including native tray/status-bar variants. (#479)

🔧 Improvements

• Preview tabs automatically reveal the active result when it moves outside the visible strip, and responsive middle truncation keeps long names recognizable without destabilizing the layout. (#478)
• Windows application and tray artwork now uses native icon sizes and dedicated small transparent variants for clearer rendering across light/dark themes and common display scales. (#484)
• Remote-compute harvests, result summaries, and notifications resolve through the active data-storage location and pause writes during data migration. (#480)
• Public project documentation now leads with clearer local-first research-workbench language, more descriptive imagery, and answer-first summaries for search and generative discovery. (#487)

🐛 Bug Fixes

• Permission previews use the correct syntax-token palette in dark mode, keeping code readable against dark surfaces. (#472)
• Codex can save through the verified app-owned artifact capability without a redundant approval prompt; ordinary writes and unverified tools remain gated. (#488)
• Provenance environment evidence now bounds retained operation history by entry count and size while preserving recovery-critical state and reporting cumulative truncation. (#489)
• Reviewer findings wrap checksum-like and other long unbroken evidence inside their cards instead of overflowing the layout. (#490)

📦 Install

Requirements: macOS 12+ (Apple Silicon or Intel), Linux x64, 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:

Platform	Package
macOS (Apple Silicon)	DMG for ARM64
macOS (Intel)	DMG for x64
Linux	AppImage or Debian package for x64
Windows	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 not notarized and may require approval through macOS Privacy & Security.

Windows — first launch (unsigned build). No Authenticode certificate yet, so SmartScreen shows a bypassable "unrecognized app" prompt (More info → Run anyway). 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

Downgrading after v0.8.0. Installing an older application does not reverse data written by a newer version. To prepare an isolated copy compatible with v0.7.3, quit Open Science and run the following command before installing the older release:

open-science rollback-to-0.7.3 --yes

The command preserves newer data and activates a separate rollback data location; it does not rewrite the current v0.8.0 state.

🧭 What's in this release (maturity)

• ✅ Implemented: a local-first desktop, localhost-web, headless, CLI, and task-SDK surface over persistent projects and sessions with selectable message branches; selectable Claude Code, OpenCode, and Codex agent frameworks; multi-provider model configuration with combined model/reasoning selection; persistent Python/R/REPL kernels and remote SSH execution; immutable artifact versions with inspectable audit evidence; multi-format previews; file-based skills; permissioned scientific connectors; opt-in review and bounded correction; configurable storage; and installers for macOS, Windows, and Linux.
• 🚧 Partial: R remains managed-only; provider choice remains constrained by the active framework's endpoint compatibility; remote compute is SSH-only; skills remain local; and review is opt-in and record-scoped.
• 🗺️ Roadmap: a unified model gateway, specialist sub-agents, Slurm/cloud-GPU execution, a public skills commons, stronger sandboxing and credential isolation, and collaborative research workflows.

🐢 Known Limitations

• R is managed-only. A bring-your-own R interpreter path is not built yet.
• Remote compute is SSH-only. Slurm, cloud GPU, and parallel sub-agent fan-out are not built yet.
• Provider choice is per framework, not one unified gateway. The available protocol depends on the selected agent backend.
• The task SDK is a first-generation surface. Task creation, polling, and artifact retrieval work; broader orchestration remains open.
• Switching agent backends cannot transfer in-flight tool state. Existing conversation history can replay, but a running action is not migrated.
• Skills are local only. There is no shared public commons, cross-machine forking, or user-facing version pinning yet.
• The reviewer is opt-in and record-scoped. It does not replace domain-specific validation of citations, units, statistics, or methods.
• Windows builds are unsigned. SmartScreen may warn on first launch; official macOS builds are notarized.
• No local GPU compute backend.
• No multi-user real-time collaboration.

🙏 Acknowledgements

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

---

What's Changed

• feat: combine model and reasoning effort pickers in composer by @roxi3906 in #468
• fix(permission): apply dark syntax token colors by @ewen-poch in #472
• fix(preview-tabs): keep active preview tabs visible by @LJLJ111 in #478
• feat(branding): replace app logo with V2 assets by @roxi3906 in #479
• feat(compute): use data root for harvested job results by @wen2zhou in #480
• feat(artifact-provenance): add durable versioned provenance by @ewen-poch in #481
• docs: refresh the release availability callout by @daanveer-tech in #483
• fix(windows): improve app and tray icon rendering by @roxi3906 in #484
• docs: improve search and generative discovery by @daanveer-tech in #487
• fix(permissions): auto-approve Codex artifact saves by @ewen-poch in #488
• fix(provenance): bound environment operation history by @ewen-poch in #489
• fix(reviewer): wrap long check content by @ewen-poch in #490
• chore(release): bump version to 0.8.0 by @ewen-poch in #492
• fix(acp): mark bounded Codex MCP test events by @ewen-poch in #494

Full Changelog: v0.7.3...v0.8.0