Appfigures for AI agents

Connect your AI assistant or agent to Appfigures and it gets live app-store data: downloads, revenue, ranks, reviews, and keywords for the App Store and Google Play, plus a catalog of millions of apps across Amazon, Steam, Roku, and more. You ask in plain English, and your agent gets the answer.

What you can ask

Here are some of the things you can ask once your agent is connected:

Your apps

• "How are downloads and revenue trending across all my apps this month versus last?"
• "How much of my revenue comes from the App Store versus Google Play?"
• "Track my subscription MRR per app, month over month."
• "How much am I spending on ads each day, and how does that line up with downloads?"
• "Which of my apps had the largest decline in downloads or revenue last month, and do the reviews hint at why?"
• "Pull this week's 1-star reviews, grouped by app version, and draft replies I can approve."
• "Where in the world are my best ratings coming from?"

App Store Optimization

• "What keywords does my app rank for in the US, and at what positions?"
• "Which of my tracked keywords have the most traffic but the least competition?"
• "How do my category ranks compare across the US, UK, and Japan right now?"
• "Am I featured anywhere at the moment, and what did being featured do to my rank?"
• "Compare my keyword rankings with my top competitor's."
• "Pull my US and Japanese store listings so I can audit the title, screenshots, and copy."

Market research

• "Rank the top 10 US iOS games by estimated revenue this month, and tell me who publishes them."
• "How many monthly downloads does an average fitness app get in the US?"
• "Find iOS apps using Firebase that earn more than $1M a month in the US."
• "Which SDKs do apps usually ship alongside OneSignal?"
• "Estimate my top competitors' monthly downloads and revenue."
• "Are new iOS games launching at the same rate as they were two years ago?"
• "How are ratings distributed across apps in my category, and how big is each store?"

Who can use this feature?

Anyone with an Appfigures account can connect their agent. Your agent gets the same access you have on your plan. For example, if your plan includes ranks, your agent can read ranks.

• Public estimates and catalog data require the Grow plan or above, with the Public Data API add-on enabled.
• Your private metrics and replying to reviews require a linked store account (App Store Connect or Google Play) on your Appfigures plan.

Requests your agent makes count against your API limits and credits, the same as any other API use. See Appfigures API, CLI, and MCP: Access, Limits, and Public Data.

Pick how to connect

There are four ways to connect. They all follow the same Appfigures access rules, so pick the one that matches where your agent runs.

Connect an AI app (MCP)

MCP (Model Context Protocol) is the standard way AI apps connect to outside tools. Appfigures runs a hosted MCP server, so there's nothing to install. Add this URL to your app, and sign in with your Appfigures account when it asks:

https://mcp.appfigures.com/mcp

Claude (web, desktop, and mobile)

1. In Claude, go to Customize > Connectors.
2. Click + Add, then Add custom connector.
3. Name it Appfigures and paste the server URL.
4. Keep the default sign-in options and click Add.
5. Sign in with your Appfigures account and approve access.

The connector then works in Claude on the web, desktop, and mobile.

On a Team or Enterprise plan, an owner adds the connector first, under Organization settings > Connectors. Members then click Connect next to it in Customize > Connectors. To see which plans include custom connectors, see Anthropic's Getting started with custom connectors.

ChatGPT

1. In ChatGPT, go to Settings > Plugins.
2. Click + Add, then Create Custom MCP Server.
3. Name it Appfigures, paste the server URL as the Server URL, and choose OAuth for authentication.
4. Click Create as a Plugin.
5. Sign in with your Appfigures account and approve access.

For screenshots, see Connecting Appfigures to ChatGPT. Custom MCP servers aren't available on every ChatGPT plan, and in Business, Enterprise, and Edu workspaces, an admin may need to allow them first. See OpenAI's Developer mode and MCP apps in ChatGPT.

Other apps

Other AI apps that support remote MCP servers may work with the same URL. Look for an option to add a custom connector or MCP server. Not every app supports the way Appfigures signs you in, so if yours can't connect, use Claude or ChatGPT.

Moving from the old setup

Earlier versions of this article set up Claude Desktop with a local server (npx -y @appfigures/cli mcp). That setup is deprecated and will be removed in a future version of the CLI. To switch to the hosted server:

1. In Claude Desktop, go to Settings > Developer > Edit Config and remove the appfigures entry.
2. Restart Claude.
3. Add the connector as described under Claude above.

Use the CLI

Use the CLI with coding agents and sandboxed agents, such as Claude Code, Codex, Cursor, or Manus. Paste this prompt to your agent:

Install @appfigures/cli from npm, then show me what it can do.

The agent installs the CLI, walks you through a one-time sign-in, and starts answering. The CLI needs Node.js 22 or later.
To skip the interactive sign-in, for example in CI, a script, or a server, give the CLI an API key instead. See Create an API key.
For the full command reference, see the Appfigures CLI README.

Build your own agent with the Agent Toolkit

If you're building an AI agent in TypeScript or JavaScript, the Agent Toolkit (@appfigures/agent-toolkit) gives it Appfigures as ready-made tools, so you don't have to write API calls yourself. It includes:

• Agent tools for the Vercel AI SDK, OpenAI, and LangChain. Your agent calls Appfigures as tools while it runs.
• A typed client for code your agent writes, such as scripts and scheduled reports.

Install it with npm install @appfigures/agent-toolkit, and set your API key as APPFIGURES_API_KEY (see Create an API key). For setup and examples, see the Agent Toolkit README. If a coding agent is doing the setup, point it at the toolkit's llms.txt.

Use the REST API

For full control, or to work in a language other than TypeScript or JavaScript, call the Appfigures API directly with an API key. See the API documentation.

Create an API key

The CLI, the Agent Toolkit, and the REST API can sign in with an API key. You need one for anything without a browser, such as CI, a script, or a server. To create one:

1. Go to appfigures.com/developers/keys.
2. Click Create a New Client.
3. In the pop-up, click Create Personal Access Token.
4. Set the token in your environment:

   export APPFIGURES_API_KEY=<your-token>

Note: The Create Personal Access Token button appears only after you create a client.

The MCP server doesn't use API keys. You sign in with your Appfigures account instead.

Changes your agent can make

Your agent acts with your Appfigures account's access. Most of what it does only reads data. A few actions make changes: replying to reviews, and tracking or untracking keywords.

• MCP: most AI apps ask you to approve these actions before they run. Keep that approval setting on.
• CLI: replying to a review or untracking a keyword won't run unless the agent adds --yes.
• Agent Toolkit: every action that makes a change refuses to run until your code approves it.

Troubleshooting

Claude doesn't show "Add custom connector." On a Team or Enterprise plan, ask an owner to add Appfigures under Organization settings > Connectors.
On the Free plan, you can add one custom connector.

Claude asks which OAuth client to use. Choose Use Claude's published identity, and leave the Client ID and Client Secret fields empty.

Updated on: 09/10/2026