Roadmap
LynxDock is built in the open and tracked as epics. This is the honest state of the platform — what's done, what's in progress, and what's planned next.
- Epic 0Completed
Foundation & Runtime
The infrastructure the whole platform is built on: a polyglot monorepo, generated protocol types, shared libraries, CI, and a publish-gated release pipeline.
- pnpm + Turborepo monorepo, Rust Cargo workspace
- @lynxdock/shared, @lynxdock/protocol (Rust ↔ TypeScript via ts-rs)
- @lynxdock/gspec with a Golden-IR contract test and browser support
- CI (JS + Rust) and Changesets release pipeline
- Epic 1Completed
Workspace & Identity
The top-level container and who you are, stored locally. Everything else hangs off the workspace.
- Local workspace (name, icon) as the root container
- Profile: display name, avatar color, Rust-side validation
- Settings applied live: theme engine, reduced motion, privacy status
- Desktop sidebar: Workspace → Profile → Settings
- Epic 2Completed
Local Messaging
A full local-first chat client — the messaging layer everything networked now builds on. Workspace → Channel → Message.
- Channels and messages: compose, edit, delete
- Reactions, @-mentions, and local file attachments
- Full-text search (SQLite FTS5) with channel and attachment filters
- Unread counts, previews, day separators, relative timestamps
Next: Now syncing over the network in Epic 3.
- Epic 3In Progress
Networking
Making messaging real-time over a self-hostable server: a WebSocket hub, live sync, and resilience on flaky connections.
- WebSocket hub transport + RPC
- Server-backed read state with live cross-device sync
- User display-name resolution over the wire
- Offline outbox: queues while disconnected, drains on reconnect
- Self-hostable server (Rust / axum) with backup, export & import
Next: Built on the protocol contracts already generated in Epic 0.
- Epic 4Planned
Communities & Servers
Friends, servers, and shared channels over the network, with membership and roles.
- Friends & presence
- Servers & membership
- Roles & permissions
- Epic 5Completed
Voice & Screen Sharing
Real-time voice and screen sharing over an SFU (LiveKit), with a clean two-plane split between control and media — delivered across five phases, hardening included.
- Control plane: call roster with join / leave / mute / screen-share state broadcast live
- Room-scoped access tokens and TURN credentials issued server-side
- Audio media: microphone publish, remote audio, active-speaker ring, deafen
- Input and output device pickers, with autoplay unblocking handled
- Screen sharing: publish, subscribe, and a multi-tile viewer
- Hardening: connection state, classified errors with retry, per-participant quality dots
Next: File transfer over the same media plane is next (Version 2.3).
- Epic 6Planned
AI & Automation
Agents and tools wired into the workspace, built on the @lynxdock/ai tool contracts.
- Tool contracts
- Workspace-aware agents
- Epic 7Planned
Plugins
Third-party extensibility via the @lynxdock/plugins SDK — a capability model and plugin host.
- Plugin SDK & manifest
- Capability model
Next: The point at which the packages begin publishing to a registry.
- Epic 8Planned
Studio
The engineering environment, including GSpec Studio, matured into a real tool for building on the platform.
- GSpec Studio
- In-browser spec validation
- Epic 9In Progress
Mission Control & Observability
One honest view of the whole system: a shared status contract, a live hub every module reports into, and an event timeline — so health, activity and progress are observable rather than guessed at.
- System status contract with a worst-wins health roll-up, generated to a single artifact
- Mission Control hub: modules, metrics and entities aggregated into one state
- Live server feed over the WebSocket hub — connections, active voice rooms, host CPU/RAM
- Commander reporting: squadrons and operations, created from the desktop app
- Event bus and timeline: retained event history replayed into a live activity log
- Studio session observer: development sessions reported into the same hub
Next: Dashboards, an integration layer, and AI Workforce views are designed and next.
Want in early?
Register interest for alpha builds and development updates, or follow the work on GitHub.