reph for ai agents
Let AI agents draft your documents. Review their work like a pull request.
Agents read and write the same markdown your team edits, through scoped tokens, and their changes wait for a person to accept them.
Start with the Agent instructions template →Sign in
Opens on reph with the template ready to start. No account needed; sign in to keep your document.
What you get
Connect Claude in a minute
Add reph as a custom connector in claude.ai, or run claude mcp add --transport http in Claude Code. Sign-in is OAuth, and revoking the token in your profile disconnects it.
Suggestions, not silent edits
Tokens are scoped read, suggest or edit. With suggest, every change an agent makes arrives as a suggestion a person accepts or rejects. Agents can never accept their own.
A review loop that closes
The agent asks for review, waits, and learns what you accepted, rejected or commented on, along with your message back to it.
Tools that check their own work
lint_doc catches broken charts, links and fences before a person sees them, and insert_diagram draws an Excalidraw canvas from a simple skeleton.
Every built-in template comes with a prompt you can paste into your agent to fill it in. Browse the templates.
Start from a template
Agent instructions
The charter your AI coding agents work from, project overview through review checklist
Prompt spec
A versioned spec for one production prompt: purpose, model and parameters, the system prompt, inputs, output contract, few-shot examples, failure modes, guardrails and history
AI feature eval plan
How you will measure an AI feature before and after launch: the task, a quality rubric, the eval dataset, metrics, baseline vs target, human review, regression gates and a cost and latency budget
Before you decide
How reph compares to the tools you already use: compare. Plans and seats: pricing. Security and data handling: trust.