Verified project record

backnotprop/plannotator

plannotator is a local, browser-based review surface for AI coding agents: users visually annotate plans, specs, markdown, HTML, and code diffs, then send structured feedback back to the agent. It plugs into coding agents via hooks and commands and supports many major agents through a single installer.

Project overview

It closes the loop between human review and coding agents: approve/deny decisions and annotations are returned to the agent as structured feedback it acts on, and one installer covers Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, and Pi.

Project type: AI Agent · AI Coding
Use cases: Coding & Development
Deployment: Refer to project documentation
License: Apache-2.0

Best for

Developers and AI engineers who use coding agents and want a visual, browser-based way to review plans, diffs, and artifacts with feedback returned to the agent.
Teams using multiple coding agents (including Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, and Pi) who want one installer to configure them all.

Key capabilities

When an agent proposes a plan via ExitPlanMode, the review surface opens in the browser; the user annotates and approves or denies, with structured feedback sent back to the agent.
Review uncommitted local changes, remote GitHub PRs and GitLab MRs in a diff viewer; comments go back to the agent; works with Git, GitButler, Jujutsu, and Perforce.
Annotate local markdown files, browse and annotate folders, or fetch and annotate any URL via slash commands or CLI.
Annotate and review rendered HTML artifacts, optionally rendering HTML as-is.
List active sessions, reopen sessions in the browser, and browse saved plan decisions read-only.
One installer auto-detects installed agents (Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, Pi) and configures hooks, skills, and slash commands.
Default-off Vim controls profile with semantic navigation, visual selection, and an optional HUD for plan and annotate views.

Getting started

Run `curl -fsSL https://plannotator.ai/install.sh | bash` (single curl/PowerShell installer). The installer auto-detects installed agents and configures hooks. Trigger agent plan mode; the browser opens with the review UI, where you annotate and approve or deny. Plan review needs no command.

Evidence and sources

README: Plannotator is a local, browser-based review surface for AI coding agents: Claude Code, Codex, Copilot CLI, Gemini CLI, OpenCode, Kiro, Droid, Amp, and Pi.
README: When the agent proposes a plan, html, or finishes writing code, the work opens in your browser and you mark it up, comment, and send feedback directly to the agent for it to act o…
README: Review local changes or remote PRs. Comment on diffs, suggest code. Your comments go back to the agent. Works with Git, GitButler, Jujutsu (`jj`), Perforce (`p4`), GitHub, and Git…
README: Annotating a rendered HTML artifact
README: brings Plannotator-style review to the terminal: annotate terminal text, review whole Markdown documents and your coding agent's replies inside Herdr, and send the feedback straig…

Source updated at: 2026-09-29T18:30:44Z

Open original GitHub project