If you're building software for real humans, humans should review the prompts powering that development. PCR.dev surfaces your AI prompt history as a structured review artifact, by project, branch, and model, so your team can learn from each other’s workflows, exchange feedback, and collaborate more effectively, much like code reviews.
Built around the way real teams work with AI.
Every prompt is scoped to a registered project and branch. Review the exact AI decisions that shaped each codebase. Nothing from other work leaks in.
Prompts aren’t just logged—they’re bundled into meaningful units of work, so your team can review workflows and give feedback like code reviews.
See which AI model made which decisions. Color-coded by family (Claude, GPT, Gemini, and more) so reviewers can evaluate the tool alongside the prompt.
Give senior engineers visibility into how the team uses AI, so process issues surface before they compound.
Captures from Cursor, Claude Code, and any MCP-compatible AI tool automatically. One platform for your whole team, regardless of which AI tools they use.
Prompts are captured as you go, with lightweight controls to bundle and organize them—no setup or workflow disruption.
No configuration files. No API keys. Just install and go.
Also available via Homebrew and as a standalone binary
Opens your browser. Create a CLI token in Settings and paste it back.
Registers the project locally and syncs it to your dashboard
Watches Cursor and Claude Code sessions in the background
Everything you need to run Prompt & Code Reviews with your team