Skip to main content
Watchfire
Back to blog

Hello from Watchfire

By Nuno Coração1 min read

Welcome to the Watchfire blog. We are building a remote control for AI coding agents: a daemon that takes clear specs, scopes them into tasks, and runs each one inside a sandboxed git worktree so the work stays reviewable. The full story, the architecture, and the install guide live in the docs, and the source is on GitHub.

This post ships alongside v7.1.0 Forge. Forge focuses on the rough edges of long sessions: the chat input no longer drops keystrokes during heavy agent output, mode switches surface a quick toast so you can confirm where you landed, and the prompt is visible the whole way through a task instead of disappearing once execution starts. Small fixes, but they make Wildfire mode feel a lot less like an experiment.

What to expect from this blog: short release notes when a new version cuts, longer pieces when we change something interesting under the hood (sandbox policy, the gRPC schema, how worktrees get reused), and the occasional behind-the-scenes look at the agent loop. No newsletter sign-up yet, no comments. If you want to follow along, the post list lives at /blog and there is an RSS feed at /blog/feed.xml.

More posts

Firestorm 9.0: Watchfire becomes a factory agents can drive

7 min

For nine majors Watchfire drove coding agents. Firestorm inverts that: watchfire mcp serve exposes the whole orchestrator to any MCP-capable client as an 18-tool factory, so an outer agent plans and reviews while Watchfire manufactures the code in sandboxed, worktree-isolated runs. It is the fourth thin client, it holds no orchestration logic of its own, and it is local-only by construction.

Inferno 8.0: one window per project

6 min

Inferno is the first feature-forward major since Beacon, and it's built around a single idea — supervising many projects at once. The Electron GUI goes multi-window, the Project View flips to chat-primary, wildfire becomes a GUI control, the dashboard becomes mission control, and a new analytics layer measures what the agents actually shipped, not just how many tasks closed.

Forge 7.3: the 300 GB log, and a quieter GUI

6 min

The 7.2 → 7.3 arc is a tidy little story about operability. We made the daemon log durable, the durable log grew to 300 GB, and 7.3 caps it. Along the way we closed two wildfire YAML bugs that were silently eating tasks, and gave the GUI a focus-chat mode. Here's what shipped.