AI Visibility API: Track How ChatGPT, Claude, Gemini, Perplexity and Grok Mention Your Brand

An AI visibility API gives you, as data, how AI assistants answer the questions your customers ask and whether your brand is in the answer. Arvow's LLM Brand Monitor API does this for ChatGPT, Claude, Gemini, Perplexity and Grok through 13 REST endpoints that any script or coding agent can call. This page covers what the API returns, every endpoint, what a prompt run costs, and how it compares with other AI visibility APIs.

What the AI visibility API returns

• Visibility and sentiment scores for the brand, overall and per AI assistant.
• Each assistant's full latest answer to every tracked prompt, with the sources it cited and the brands it mentioned.
• Prompt gaps: each prompt is marked covered, competitor gap (competitors appear, you do not) or visibility gap.
• AI Mentions: the publisher pages the assistants cite, ranked as outreach opportunities, with the page type (roundup, directory, review, editorial or community), a contact route and a suggested pitch.

Prompts are also classified by type (organic, brand specific or competitor comparison) and by intent (branded, informational, navigational, commercial or transactional) after their first run.

LLM visibility API endpoints

All paths start with /api/v0.1/projects/{projectId} on api.arvow.com. The full schema is in the API reference.

Request	What it does	Cost
GET /monitors	List brand monitors with their average visibility and sentiment	Free
POST /monitors	Create a monitor and start running its prompts	6 credits per prompt run
POST /monitors/extract-brand	Read a website and suggest the brand name, aliases, description and starter prompts	Free
GET /monitors/{monitorId}	Brand settings and current scores, overall and per assistant	Free
PATCH /monitors/{monitorId}	Change settings, or pause and resume automatic runs	Free
DELETE /monitors/{monitorId}	Delete the monitor with its prompts and results	Free
GET /monitors/{monitorId}/prompts	Tracked prompts with scores and mentioned brands per assistant	Free
POST /monitors/{monitorId}/prompts	Add 1 to 100 prompts and run the new ones	6 credits per prompt run
DELETE /monitors/{monitorId}/prompts	Stop tracking selected prompts	Free
GET /monitors/{monitorId}/prompts/{promptId}	One prompt with each assistant's full answer and cited sources	Free
POST /monitors/{monitorId}/prompts/run	Ask every assistant the selected prompts again	6 credits per prompt run
GET /monitors/{monitorId}/mentions	Prompt gaps and the publisher pages assistants cite	Free
PATCH /monitors/{monitorId}/opportunities/{opportunityId}	Move a publisher through outreach: new, shortlisted, contacted, in discussion, secured or not fit	Free

Quick start: track a brand in four calls

Every request carries your organization API key as a Bearer token: Authorization: Bearer YOUR_API_KEY.

1. Find your project. GET /api/v0.1/projects returns your projects and their ids.
2. Create the monitor. POST /api/v0.1/projects/{projectId}/monitors with a JSON body such as {"brandName": "Acme", "brandUrl": "https://acme.com", "prompts": ["Best invoicing software for freelancers"]}. Only brandName is required, and the prompts start running right away.
3. Read the answers. Poll GET .../monitors/{monitorId}/prompts until isRunning is false. Each prompt then carries a visibility score, a sentiment score and the latest answer per assistant.
4. Find the gaps. GET .../monitors/{monitorId}/mentions lists the prompts that miss your brand and the pages the assistants cited instead.

If you do not have prompts yet, call POST .../monitors/extract-brand first: it reads your site and suggests starter prompts. Our guide to getting recommended by AI covers what to track next.

Using the AI visibility API from an agent

You do not have to write the calls yourself. Arvow's MCP server exposes the same 13 brand monitor operations as tools, so ChatGPT, Claude or a coding agent such as Claude Code or Codex can do the work in plain language:

• "Create a brand monitor for acme.com and track these ten prompts."
• "Which of our prompts mention competitors but not us?"
• "List the pages AI assistants cite for our prompts and mark the three best as shortlisted."

ChatGPT connects to the MCP server with OAuth sign-in. Claude, Claude Code and Codex connect with your API key, and the help article has the steps. An agent can also call the REST API directly with the same key.

What the AI visibility API costs

Every endpoint is free to call. What uses credits is running a prompt: 6 credits per prompt run, exactly like the dashboard, and one run asks all five assistants. While monitoring is on, each prompt re-runs about every 3 days. As a rough guide, 20 prompts re-run every 3 days is about 200 prompt runs a month, or about 1,200 credits.

Brand monitors are part of the Business plan and above: 3,000 credits a month, from $69 a month billed yearly or $169 billed monthly, on arvow.com/pricing. A Solo API key can use the rest of the API, but it cannot create or run brand monitors.

Limits to know before you build

• Five assistants: ChatGPT, Claude, Gemini, Perplexity and Grok. Google AI Overviews, AI Mode and Microsoft Copilot are not in the list.
• Scores start empty. Visibility and sentiment are null until the first run finishes.
• One change at a time. Updates fail with ApiMonitorBusy while a monitor is running prompts or updating AI Mentions; retry when the mentions endpoint reports the pipeline is complete.
• Request size: up to 100 prompts per request, each up to 1,000 characters.
• Renaming resets results. Changing the brand name, URL or aliases clears the monitor's results until prompts run again.

Other AI visibility APIs

Keyword.com's AI Visibility API covers ChatGPT, Perplexity, Gemini, Google AI Overviews, Google AI Mode and Microsoft Copilot, and says API usage is based on your plan with no per-call fees. LLM Pulse's comparison, which ranks its own API first, reports that Profound's API is Enterprise-only, that Otterly includes API access on its middle and upper tiers, and that SE Ranking includes API and MCP access across its plans.

What is different about Arvow is what sits next to the data. The same API key generates and publishes articles to your CMS, and AI Mentions turns every cited page into an outreach task with a status, so an agent can go from "we are missing from this answer" to doing something about it.

FAQ

What is an AI visibility API?

An API that runs customer-style prompts against AI assistants on a schedule and returns whether a brand was mentioned, how it was described, and which sources the answer cited.

Which AI assistants does Arvow's AI visibility API cover?

ChatGPT, Claude, Gemini, Perplexity and Grok. Scores are returned overall and per assistant.

Is there an MCP server for LLM visibility tracking?

Yes. Arvow's MCP server includes the 13 LLM Brand Monitor tools: create a monitor, add and run prompts, read visibility and sentiment scores, and list AI Mentions. It connects to ChatGPT, Claude, Claude Code and Codex, and brand monitors need the Business plan or above.

How much does the AI visibility API cost?

The endpoints are free. Each prompt run costs 6 credits, taken from your plan's monthly credits. Brand monitors need the Business plan or above.