Plannotator is a local, browser-based review surface for AI coding agents that lets a developer annotate plans, documents, HTML artifacts, and code diffs, share them with a team, and send feedback back to the agent in one action.

What it is

Plannotator is an open-source TypeScript project, licensed under Apache-2.0, that plugs directly into AI coding agents through their hooks and commands. The README names Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, and Pi as supported agents. When the agent proposes a plan, produces an HTML artifact, or finishes writing code, that work opens in the browser, where the reviewer marks it up, leaves comments, and sends the feedback to the agent for it to act on.

The concrete problem it solves is the review loop around agent output. Instead of reading a plan as raw terminal text and replying with a freeform message, the plan, spec, message, or HTML opens as an annotatable document through commands such as /plannotator-annotate and /plannotator-last. The same surface extends to code: local uncommitted changes, remote pull requests, and merge requests can be reviewed diff by diff, with comments routed back to the agent. Version control coverage includes Git, GitButler, Jujutsu (jj), Perforce (p4), GitHub, and GitLab.

Key capabilities

• Annotate review targets from several sources: a local markdown file (/plannotator-annotate README.md), a folder (src/), any URL, or an HTML report with --render-html to render it as-is rather than converting it.
• Plan mode runs with no command at all: it is wired in through each harness's hooks, so the markdown review surface opens whenever the agent creates a plan.
• Code review through /plannotator-review for uncommitted changes, GitHub pull requests, and GitLab merge requests, plus plannotator review --gitbutler for an active GitButler workspace, reviewable as a whole workspace, a single stack, or one branch layer.
• AI built into the review surface: ask AI about anything under review, or launch AI reviews that post comments directly to the diff.
• CLI session and archive handling: plannotator sessions, plannotator sessions --open 1 to reopen a session in the browser, and plannotator archive to browse saved plan decisions read-only.
• Herdr Annotate brings the same review style to the terminal, annotating terminal text and agent replies inside Herdr and sending feedback back as the agent's next message.
• Codex support is expressed differently: !plannotator … in place of slash commands, or the $plannotator-* skills.

Who uses it and how

• Developers running plan mode in Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, or Pi, who get a review surface automatically when a plan is created.
• Reviewers working on GitHub pull requests and GitLab merge requests, who comment on diffs and route those comments back to the agent.
• GitButler users reviewing an entire workspace, one stack, or a single branch layer rather than a flat diff.
• Terminal-centric users who install Herdr Annotate and keep working inside Herdr; annotations are saved in the Plannotator data directory, so terminal reviews and app reviews compound.
• The repository also carries an obsidian topic alongside agents, claude-code, code-review, codex, opencode, pi-mono, plan-mode, and skills.

Getting started

Installation follows the agent's hooks and commands as described in the installation guide on the official site, after which the review commands such as /plannotator-annotate and /plannotator-review become available. For terminal use, herdr plugin install plannotator/herdr-annotate adds the Herdr plugin, and brew install plannotator/tap/plannotator-tui installs the standalone TUI for document review without Herdr.

How it compares

This registry does not list paid products that Plannotator replaces, and no directly comparable third-party tools are named in the material provided, so on the evidence here it stands alone. The nearest neighbours are its own siblings: Herdr Annotate, which brings the same review model into the terminal, and Plannotator TUI, which powers document review standalone and shares the same data directory.

When to use it — and when not to

A self-hoster runs a local browser-based review surface for each agent harness; no database, object storage, or SMTP service is mentioned as a requirement. Plannotator does not collect usage telemetry or analytics, and plans, diffs, annotations, drafts, history, and configuration stay local by default, but each review, annotate, archive, share-portal, and code-review surface checks GitHub for the latest release on load, and there is currently no opt-out setting; local Git code review may also query the configured origin with git ls-remote. Teams that require zero outbound requests, or that want a hosted multi-tenant service rather than a local tool, should look elsewhere, and the excerpt's thin coverage of standalone installation plus 155 open issues are worth weighing before a wide rollout.

herdr plugin install plannotator/herdr-annotate

/plannotator-annotate README.md                  # Local markdown file
/plannotator-annotate src/                       # Browse and annotate files in a folder
/plannotator-annotate https://docs.rs/…          # Fetch and annotate any URL
/plannotator-annotate report.html --render-html  # Render HTML as-is instead of converting
/plannotator-last                                # Annotate the agent's last message

/plannotator-review                    # Review uncommitted changes
/plannotator-review <github-pr-url>    # Review a GitHub pull request
/plannotator-review <gitlab-mr-url>    # Review a GitLab merge request
plannotator review --gitbutler         # Review an active GitButler workspace

plannotator sessions                   # List active Plannotator sessions
plannotator sessions --open 1          # Reopen a session in the browser
plannotator archive                    # Browse saved plan decisions read-only