Quickstart

Five steps, each one curl command. By the end you will have approved a card and had a conversation with Revnu from a terminal.

No account yet? npx @revnu/cli setup creates one and sets it up from your terminal. See Install the CLI.

1. Mint a token

In the dashboard, open Settings → Developers, give it a label and only the scopes the tool needs. The token is shown once. We store a hash. Export it for the rest of this page:

export REVNU_API_TOKEN=rvp_…

2. Check who you are

GET /me needs no scope, so it is the call to verify any token with.

curl -H "Authorization: Bearer $REVNU_API_TOKEN" https://revnu.com/api/v1/me

{
  "me": {
    "label": "quickstart",
    "scopes": ["review:read", "review:write", "agent:message"],
    "account": { "id": "…", "name": "Acme" },
    "inSetup": false,
    "limits": { "reads": { "perMinute": 300, "burst": 600 }, "…": "…" }
  }
}

3. Read what is waiting

curl -H "Authorization: Bearer $REVNU_API_TOKEN" "https://revnu.com/api/v1/review?limit=10"

Every card carries actions, the exact verbs the API accepts for it, and primaryLabel, what the dashboard's main button says. Drive your UI from those, never from the card type.

4. Approve it

curl -X POST -H "Authorization: Bearer $REVNU_API_TOKEN" \
  -H "content-type: application/json" -d '{"action":"approve"}' \
  https://revnu.com/api/v1/review/k97…/answer

5. Ask the agent something

curl -X POST -H "Authorization: Bearer $REVNU_API_TOKEN" \
  -H "content-type: application/json" \
  -d '{"text":"What did we ship this week?","waitSeconds":90}' \
  https://revnu.com/api/v1/agent/messages

With waitSeconds the request stays open until Revnu replies, and the response is 200 with the reply inline. Without it you get 202 and a requestId to poll. A new message stops work Revnu is still doing for this token, so wait for the reply before sending the next one, or see Talk to Your Agent for handing over several jobs.

Next

• Skip the curl: npm install -g @revnu/cli && revnu login. See Install the CLI.

• Use Revnu from Claude, ChatGPT or Cursor: MCP Server Setup.

• Put the queue on a screen: Build a Review Screen.