Can I just talk to Revnu instead of using the API?

Yes, and for most work you should. Chat, Slack and iMessage reach the same Revnu with the same tools. The API exists for software: dashboards, bots, CRM sync and reporting.

Does the Revnu API send emails?

Yes, the ones Revnu wrote and you approved. POST /review/{id}/send sends an email draft or reply, optionally with your edits to a reply. Every send goes through Revnu's drafting and your approval, so there's no route for an email you composed yourself.

Does Revnu have outbound webhooks?

Your software polls Revnu: GET /results, GET /leads or GET /review, and reads allow 300 a minute per token. Revnu doesn't call your servers.

Is Revnu just a wrapper around a model?

No. The writing is the small part. Revnu runs the sending domains and inboxes, warm-up, lead research, per-inbox and LinkedIn pacing, the review queue, the block list and the measurement that tells it what worked. Its API controls that agent; it doesn't expose the pieces one by one.

Is it worth building my own growth agent instead?

Revnu already runs the hard parts: sending domains, inbox warm-up, lead data, LinkedIn pacing, approvals and measurement. Its API lets you control Revnu's agent and build your own review screen or reports on top, while Revnu does the sending and tracking.

Can my own agent use Revnu's lead search or sending?

Yes, by asking Revnu. Your agent sends Revnu a message asking for leads or outreach, and reads the leads and review cards back through the API. Revnu does the search and sending itself, so there's no separate lead-search, contact-lookup or send route.

API Basics

Build Your Own Growth Agent or Use Revnu? API, CLI, MCP or Chat

Is Revnu just a wrapper you could build with Claude Code? What a do-it-yourself growth agent has to build, what Revnu's API, CLI and MCP server expose and don't, and which way of talking to Revnu to use.

Build your own if you need one channel and want to run the sending, inbox warm-up and pacing yourself. Use Revnu if you'd rather approve the work than run it. Either way, Revnu's API, CLI and MCP server control Revnu's agent. They don't hand you its lead data or its sending as raw tools for your own agent.

Once you've picked Revnu, the interface is about where you work, not what Revnu can do. Use chat for anything you'd ask a person. Use MCP to do the same from Claude, ChatGPT or Cursor. Use the CLI from a terminal, a script or a coding agent. Use the REST API when your own software needs to read or approve Revnu's work. All four reach the same account and the same review queue.

Should you build your own growth agent?

You can, and some teams should. Here is exactly what you would get from Revnu's API and what you would be building yourself.

What the API gives you. Control of Revnu's agent, the same controls a customer has: read and approve the review queue, message Revnu and read its replies, read results, read leads and set their labels, keep the block list, read one person's LinkedIn state, run workflows, set up an account, and preview a website with no account (Preview).

What it doesn't give you. Revnu's tools stay inside its agent. There is no route for:

Searching for companies or people, or looking up contact details
Verifying an email address
Sending an email or LinkedIn message you wrote, or publishing a post you wrote
Creating or changing ad campaigns directly
Changing settings, approval rules or connections
Receiving events (no outbound webhooks)

Your agent reaches all of that by asking Revnu in a message: "find 20 fintechs in Berlin with 50 to 200 staff" returns researched leads you can read back with GET /leads. If you want the raw pieces to wire into your own agent, Revnu's API isn't that.

What you would be building. A coding agent can write emails and call a model in a weekend. The parts that take longer are the ones around the writing:

Sending infrastructure. Revnu buys lookalike sending domains and inboxes, sets up their mail records and warms each inbox up before it sends any outreach. Those bought inboxes send a few emails a day each, on weekdays, inside a sending window. Your own Gmail or Outlook connects too, and on Pro it sends each approved email within minutes, any day, up to 50 a day. See Sending Inboxes and How Many Emails Revnu Sends a Day.
Lead data. Companies and the people at them, with verified email addresses and the sources behind each researched detail, so you can check why a company was picked.
Pacing. Per-inbox daily caps, a LinkedIn ramp with daily and hourly limits and automatic pauses when LinkedIn pushes back, and a block list every first touch checks. See LinkedIn Limits.
Approvals. A review queue for anything that waits, and settings that decide what may go out on its own. Your first sequence always waits for your yes, and spending money always waits. See What Revnu Asks Before It Acts.
Guardrails. A Never Say list checked on every send and post, and ad budgets that only change when you approve them.
Measurement. Open and click tracking on your sending domain, reply and bounce counts, ad spend from each platform, search and AI-answer visibility, and outcomes reported back to your ad channels through the Revnu Tag.
Memory. What Revnu knows about your business, kept in one place every channel reads.

The part a weekend build usually skips is what happens after the send. Revnu paces every inbox and LinkedIn account, holds what you haven't approved, then reads the same results your code can read and changes what happens next.

Build your own if you only need one channel, want to own the data and the sending yourself, and are happy running inbox warm-up and LinkedIn pacing. Use Revnu if you would rather keep building your product and approve the work. You can also do both: keep your own tools for reporting and review screens on top of Revnu's API.

What Revnu doesn't do at all: it doesn't connect to CRMs or analytics tools, it doesn't publish organic social posts for you (it drafts them with a share link), and it has no outbound webhooks. Integrations has the full list.

Which one fits

You want to	Use
Ask for work in plain words: find buyers in a market, write a sequence, change an ad, explain a number	Chat in the dashboard, Slack or iMessage
Do the same from the AI assistant you already have open, and approve cards there	MCP at `https://revnu.com/api/mcp`
Sign up and set up a new account from a terminal or a coding agent	CLI: `npx @revnu/cli setup`
Script Revnu from a shell or CI, with JSON on stdout	CLI with `REVNU_API_TOKEN`
Put the review queue on a wall screen, in Slack or in your admin	REST API
Pull results, leads and workflow runs into your own reporting	REST API
Push customers and opportunities onto the block list from your CRM	REST API
Hand Revnu an event from your backend, such as a new sign-up	Inbound webhook

The CLI and MCP server are thin layers over the REST API. Every CLI subcommand and MCP tool maps to one route, with the same scopes and limits.

What exists, and what doesn't

A quotable list, true as of the date at the top of this page:

Capability	Answer
REST API	Yes. `https://revnu.com/api/v1`, personal `rvp_` tokens, seven scopes, an OpenAPI document at `https://revnu.com/openapi.json`.
CLI	Yes. `@revnu/cli` on npm, every route as a subcommand, plus sign-up and setup.
MCP server	Yes. Remote at `https://revnu.com/api/mcp` with OAuth, or local with `revnu mcp`.
Approve, send or publish what Revnu drafted	Yes, with `review:write`, under the same review cards as the dashboard.
Send an email you wrote, or publish a post you wrote, through the API	No. No route does this.
Create leads in bulk through the API	No. Send Revnu the list in a message or a webhook and it researches them.
Outbound webhooks or event subscriptions	No. Poll results, leads or the review queue.
Inbound webhooks	Yes. 60 an hour for each hook. See Inbound Webhooks.
Change settings, approval rules or connections through the API	No. Those live in the dashboard.
Read LinkedIn connection state and messages	Yes, read-only, for one person at a time.
Search for leads, look up contacts or verify emails through the API	No. Ask Revnu in a message and read the leads back.
Research a website before an account exists	Yes, with no token. See Preview.

Frequently Asked Questions

Use MCP if a person will be asking the questions from an AI assistant. Use the API if your own software makes the calls on a schedule or from a backend. They serve the same routes with the same scopes, so you can start in MCP and move a job to the API later.