Firestorm turns Watchfire inside out: instead of only driving coding agents, Watchfire is now driven by them. watchfire mcp serve exposes the whole orchestrator to any MCP-capable client — Claude Code, Codex, Gemini CLI, opencode, Copilot CLI, or a custom agent — as an 18-tool factory. The outer agent plans and reviews; Watchfire manufactures the code in sandboxed, git-worktree-isolated runs and merges the results. The canonical loop is create_task → run_task → wait_for_task → get_task → get_task_diff → iterate.
Roadmap.
What we’ve shipped, what we’re building now, and what’s on deck. Watchfire is open source, so everything here is also where you can help.
Now shipping.
The most recent releases — pulled straight from the changelog so this stays current automatically.
Inferno is the parallel-workspaces tentpole — the first feature-forward major since v4 "Beacon", built for supervising many projects at once. The Electron GUI goes multi-window: a main-process window registry replaces the single mainWindow, opening one independent window per project (single-instance lock, per-window state + session restore, Cmd+N and window cycling, IPC fan-out, and per-window PTY routing so a terminal's bytes and OS notifications never cross windows). Each project window flips to a chat-primary layout with the agent terminal as the wide left pane and Tasks/Definition/Insights/Secrets/Trash/Settings as a right reference region. The plain markdown <textarea>s become a real CodeMirror rich editor (formatting toolbar + source ⇄ split ⇄ preview, closing #22), wildfire mode lands in the GUI behind a confirm-gated start with a live Execute → Refine → Generate phase indicator, and the promoted Dashboard becomes mission control — a live home window with per-project wildfire phase, cross-project needs-attention with click-through, and a stretch always-on-top mini-monitor. Finally, code-output analytics measure what the agents actually shipped: per-task <n>.metrics.yaml captures commits/files/lines/merge-kind, rolled up into project and fleet Insights, the GUI's KPI cards and churn-by-day chart, mission-control "shipped" lines, and the CSV/Markdown exports plus weekly digest.
Forge 7.4 closes a class of "wildfire stops even though there's a ready task sitting right there, and drops back to chat" reports. Traced live on a real project, the symptom turned out to be three independent daemon bugs plus a runaway log that buried the evidence: Generate produced a ready task, the chain stamped it started, but no agent ever ran it — and the one log line that would have explained why was drowned under hundreds of MB of self-referential watcher spam. This release fixes the runaway log at the source, moves the verbose per-project trail into per-project log files so the global daemon.log stays readable, makes a launch failure unable to strand a ready task, and stops the issue detector from false-positiving on ordinary agent output.
Forge 7.3 is primarily a GUI release. The Electron app gains a focus-chat mode that collapses the center column so the right panel (Chat / Branches / Logs) takes the full row, and the running Watchfire version now sits under the sidebar logo so you no longer have to dig through Settings → About to know what you're on. The version bump to 7.3.0 lands across version.json, gui/package.json, and gui/package-lock.json. One daemon-side fix closes the size-cap deferral v7.2.1 left open: ~/.watchfire/daemon.log is now bounded after a user's log grew to 300 GB.
Browse the full release history in the changelog.
In progress.
Actively in flight. These are the pieces being shaped, prototyped, or merged behind a flag right now.
Windows daemon support
Bring `watchfired` to Windows with native service installation, signed binaries, and a Windows-shaped sandbox replacement for `sandbox-exec`. macOS and Linux already ship; Windows is the last gap.
Per-task cost tracking in the GUI
Surface live token spend and cumulative cost per task in the Electron GUI, broken down by agent backend. Backed by transcripts already written to disk — no extra telemetry.
OpenAI Codex CLI backend stabilization
Lock in Codex CLI as a first-class backend alongside Claude Code: tighter error surfaces, consistent transcript layout, parity with the chat/task/start-all modes.
Editor-mode TUI keybindings
Vim and Emacs profiles for the TUI — modal navigation, named registers for task IDs, customisable bindings via `~/.config/watchfire/keys.toml`.
Beacon v2 protocol
Second iteration of the daemon's two-way integration surface: streaming task updates, scoped capability tokens, and a stable wire format for third-party clients.
Docs search powered by the local daemon
Replace the static search index on `/docs` with a daemon-served endpoint so search results reflect the version of Watchfire you actually have installed.
On deck.
The next quarter's intended pickups. Order is rough and priorities shift — `lib/roadmap.ts` is the source of truth.
Linux GUI builds
Ship `Watchfire.app` as a signed AppImage and a `.deb` so Linux users get the same multi-project view macOS already has.
Per-project secrets vault
First-class secrets storage scoped to a project — encrypted at rest, surfaced to agents only inside the worktree, never leaked into transcripts.
Cron-triggered task runs
Schedule recurring tasks natively in `project.yaml` (`schedule: "0 9 * * 1"`) so the daemon fires drafts on its own — no external runner required.
Multi-user daemon mode
Let `watchfired` accept connections from more than one operator with per-user identity, audit log, and project-level ACLs. The piece teams keep asking for.
Marketplace for community task templates
A browsable index of community-contributed task templates, installable with one command — built on the same YAML schema as `/templates`.
JetBrains and Zed integrations
Editor plugins that mirror the VS Code workflow: open a task, jump to its worktree, watch the transcript inline.
How to help.
Watchfire is Apache-2.0 open source. Every item on this page maps to a GitHub label — the easiest entry points are tagged `good-first-issue` and `help-wanted`.
Not sure where to start? Drop into GitHub Discussions — design conversations and rough ideas land there before they become issues.