# NeuroPage developer docs

> Make a personalized landing page for every B2B lead, from an AI agent or from code.

NeuroPage makes a personalized outbound landing page for every B2B lead you give it, from nothing more than their LinkedIn profile URL. Each page is written about that lead's company and role, styled in your brand, and hosted at a permanent URL you can paste into Instantly, HeyReach, Lemlist or any sequence. Afterwards you can see who opened their page, how far they scrolled, and who clicked or booked.

You can use it from an AI agent through the MCP server, or from code and tools like Clay, n8n and Make through the REST API. Both use the same objects: a workspace, its brands and templates, the jobs you start and the assets (pages) they make.

## Start here

- **Use it from an AI agent.** Connect claude.ai, Claude Code, Claude Desktop, ChatGPT, Cursor, Grok or VS Code to the MCP server and ask it to make pages. Sign in with your NeuroPage account; no key needed. [Connect your AI agent](https://docs.neuropage.io/mcp)
- **Call the REST API.** Create an API key, make pages for 3 leads with curl, wait for the job and list the page URLs. [Quickstart](https://docs.neuropage.io/quickstart)
- **Just exploring.** Read how workspaces, brands, templates, jobs and assets fit together, or browse every endpoint. Reads work on any workspace, with or without a plan. [Concepts](https://docs.neuropage.io/concepts) | [API reference](https://docs.neuropage.io/reference) | [Contact us for sandbox access](https://neuropage.io/contact)

## Which one should I use?

| If you want to | Use | Good to know |
|---|---|---|
| Set NeuroPage up and run campaigns from a chat | [MCP](https://docs.neuropage.io/mcp) | Signs in as you (OAuth) or uses an API key. The campaign tools (leads, research, messages, Instantly) are MCP only for now. |
| Make pages from your own code, Clay, n8n, Make or a script | [REST API](https://docs.neuropage.io/quickstart) | One API key per workspace, sent as `Authorization: Bearer np_live_...`. Up to 25 leads per job; follow a job with `GET /jobs/{id}?wait=50`. |
| Be told when a page is ready or a lead acts on it | [Webhooks](https://docs.neuropage.io/webhooks) | NeuroPage POSTs to your URL for `page.ready`, `job.failed`, `lead.ready` and page activity. Set up in the app or with `subscribe_events` over MCP; there is no REST `/webhooks` endpoint. |

Reads work on any workspace. Writes (creating brands or templates, generating pages) need a paid plan or a trial with a card on file; without one they return `403 plan_required`. See [Sandbox and plans](https://docs.neuropage.io/sandbox-and-plans).

## Common ways to automate

- **Sequencers (Instantly, HeyReach, Lemlist).** Start a job for a list of leads, wait for it, then paste each lead's permanent page URL into the sequence. Over MCP, `send_to_instantly` adds leads to an Instantly campaign with their page URL and emails, and `import_from_instantly` pulls an Instantly campaign or list into a NeuroPage campaign.
- **Clay, n8n, Make and other no-code tools.** `POST /jobs` from an HTTP step, one row per lead, with the `campaign_id` of a campaign whose Send stage is a webhook. Each page URL comes back as a `lead.ready` event, one flat JSON object per lead. The [Webhooks](https://docs.neuropage.io/webhooks) page has the Clay recipe.
- **Bulk lists.** Send jobs of up to 25 leads with an `Idempotency-Key`, at most 2 running page jobs per workspace; new pages are capped at 200 per hour and 1,000 per day. A `429` carries `Retry-After`, and `dry_run: true` quotes a job before anything is charged. See [Credits and limits](https://docs.neuropage.io/credits-and-limits).
- **An agent that does it all.** From claude.ai or Claude Code, ask for a brand from your website, pages for 3 LinkedIn profiles, and who viewed them. The agent tells you what the pages will cost before it spends anything. See [Connect your AI agent](https://docs.neuropage.io/mcp).

## Good to know

- One API key belongs to one workspace, and one workspace is one client. Keys start with `np_live_` and are shown once. See [Authentication](https://docs.neuropage.io/authentication).
- A brand is the look (logo, colours, fonts); a template is the layout (sections and call to action). The API keeps `brand` and `template` separate, so one look can be reused across layouts or changed per job.
- Every job is quoted lead by lead before it starts, and each page is charged exactly its quote. A LinkedIn URL that already has a page in the same campaign is reused at no charge. Every cost and refund is in [Credits and limits](https://docs.neuropage.io/credits-and-limits).
- Pages are hosted at a permanent URL that does not expire.
- Errors are `application/problem+json` with a stable `code`, a `request_id` and, when there is one, a `fix` URL or tool. [Errors](https://docs.neuropage.io/errors) lists every code.
- Ids say what they are: `brd_` brand, `tpl_` template, `cmp_` campaign, `job_` job, `ast_` asset.

## Reference

- [API reference](https://docs.neuropage.io/reference): every endpoint with request and response examples, rendered from the OpenAPI document.
- [OpenAPI 3.1](https://neuropage.io/api/v1/openapi.json): the REST API as `openapi.json`, webhook payloads included.
- [Markdown for agents](https://docs.neuropage.io/reference.md): the whole guide and reference in one Markdown file. Every page here is also Markdown: request it with `Accept: text/markdown`, or open [its Markdown URL](https://docs.neuropage.io/index.md).
- [llms.txt](https://neuropage.io/llms.txt): a plain-text map of the site for language models.
- [API and MCP changelog](https://docs.neuropage.io/changelog): what changed in the API and the MCP server, with an [RSS feed](https://docs.neuropage.io/changelog/rss.xml).
- [Product changelog](https://neuropage.io/changelog): what changed in NeuroPage itself.
