Skip to content
Audience Intel
Developers

The engine, in your own tooling.

Market scans, report facts, review reads, validation studies and more, through a versioned REST API and a remote MCP server your assistant tools can connect to. Every key carries explicit scopes and a daily spend cap.

Offered · run by a specialist Live, sales-led. Keys are issued by us after a conversation; there is no self-serve sign-up.

The developer docs' MCP quickstart: the remote server address, the command that adds it to Claude Code, and the list of tools it exposes
Captured from /developers
In the OpenAPI 3.1 document, checked against shipped routes
42 paths
Source: src/lib/api/openapi.ts · checked 2026-10-13
Counted from the tool list in code
19 MCP tools
Source: src/lib/mcp/tool-names.ts MCP_TOOL_COUNT · checked 2026-10-13
Per key; over the cap is refused
Daily spend cap
Source: src/lib/api/spend.ts · checked 2026-10-13
Within v1: no renames, no removals
Additive only
Source: src/lib/api/openapi.ts info.description · checked 2026-10-13
What it is

What it is, and how it works.

The Audience Intel engine in your own tooling: a versioned REST API and a remote MCP server, with per-key scopes and daily spend caps. Keys are issued after a conversation.

Everything the site does runs on one engine, and the engine has an API. Build your own dashboard, internal workflow or assistant set-up, and call the engine directly instead of through our pages.

The REST API lives under one versioned path. It covers market-scan reports and the facts inside them, review reads, validation studies, prospecting runs, continuous assessment and more, all described in an OpenAPI 3.1 document you can download. Within version one, changes are additive only: new fields and endpoints, never renamed or removed ones, and never a new required input. Errors share one envelope with stable codes.

The MCP server exposes the same engine as tools an assistant client can call: running and reading scans, querying and correcting report facts, review reads, the synthetic panel, prospecting, assessment and outreach checks. It authenticates with the same keys.

Keys are issued by us after a conversation, not through a sign-up form. Each key carries explicit scopes (read reports, write scans and so on) and a daily spend cap in pence. A request that would exceed the cap is refused with a clear error. A key with a cap of zero cannot use anything that spends. An API that can run paid work needs a ceiling agreed in advance.

Who it is for

  • Agencies building reporting on top of the engine
  • Teams with their own dashboards or internal tools
  • Developers connecting assistant clients over MCP
What it does

What it reads, and what you see.

Versioned REST API

An OpenAPI 3.1 document, one error envelope and additive-only changes within v1.

Remote MCP server

The engine's main jobs as tools an MCP client can call, with the same keys.

Scopes and spend caps

Every key scoped to what it needs, with a daily spend ceiling.

By hand, or with it

The same job, two ways.

By hand

  • Copy results out of pages by hand
  • Rebuild research pipelines yourself
  • Give a tool open-ended spending power
  • Chase breaking API changes

With API and MCP server

  • The engine called from your own code
  • Assistant clients connected over MCP
  • Scoped keys with a daily cap
  • A stable, additive-only v1
What you walk away with

What you keep.

Programmatic access

Scans, facts and studies from your own tooling.

An MCP connection

The engine available to assistant clients.

Controlled spend

A daily ceiling per key, agreed up front.

A stable contract

A published spec and additive-only changes.

How it works

Three steps.

  1. Talk to us

    Tell us what you want to build and which parts of the engine you need.

  2. Receive a scoped key

    We issue a key with the scopes and daily cap you need.

  3. Build

    Call the API or connect an MCP client, using the published spec.

Limits

What it does not do.

  • There is no self-serve sign-up; keys are issued after a conversation.
  • Each key is limited to its scopes and its daily spend cap.
  • Some engine functions are internal and not part of the public scopes.
  • Results carry the same limits as the products they come from. An estimate through the API is still an estimate.
Price

Enterprise, price on application.

Keys and pricing are agreed in conversation; the Enterprise tier on the pricing page covers this.

See pricing
FAQ

API and MCP server: questions

Can I get a key today?

Yes, after a conversation. Keys are issued by us, not through a sign-up form.

Where is the specification?

The OpenAPI 3.1 document is published, and the developer page links to it.

What stops a key running up a bill?

A daily spend cap on every key. A request that would exceed it is refused, and a key with a cap of zero cannot use anything that spends.

Will v1 change under me?

Only additively: new fields, endpoints and values, never renames, removals or new required inputs.

Run it on your own business.

Live, sales-led. Keys are issued by us after a conversation; there is no self-serve sign-up.