A human check on AI text

Agents now send email, open pull requests and reply to clients. Most of the time the text is fine, and when it is not, it still goes out under your name. reright puts a person between the agent and the reader. With "Check before it's used" on, every rewrite waits for you.

How the check works

Turn on "Check before it's used" in Settings, under Rewriting. Each time the agent calls reright, the rewrite waits in Rewrites and the agent gets a review link, which it gives you once before it waits. You approve, edit or reject. What you approve is exactly the text in the editor, character for character. If you reject, the agent is told to stop and wait for your instructions.

Reviewing quickly

The review page shows the kind of text, where it goes and which agent wrote it. You can switch to the agent's own text or show the changes. Ctrl+Enter approves, and the next waiting rewrite loads. Push notifications on your browser or phone tell you when one is waiting.

What the agent cannot skip

In Claude Code, a hook blocks git commits in shared repos, gh PR and issue text and Gmail sends unless that exact text came back from reright in the last 24 hours. A reject locks the session until you type something. Chat assistants such as Claude.ai and ChatGPT have no hook and follow the instructions you give them. The docs list what gets past the hook.

Your edits teach it

When you edit a rewrite before you approve it, reright keeps your version to learn from. Turning the check on for a while is the fastest way to teach it how you write. Later you can turn it off and let rewrites go through on their own, with the fact checks still in place.

Try it on the next thing your agent writes. The free plan needs no credit card.