AppFollow MCP and AI Toolkit

Article author
AppFollow Team
  • Updated

AppFollow provides two ways to work with your AppFollow data using AI:

  • AppFollow MCP Server — connects AppFollow to MCP-compatible AI assistants such as Claude, ChatGPT, Codex, and Cursor, allowing them to access your live AppFollow data.
  • AppFollow AI Toolkit — an open-source plugin built on top of MCP that provides ready-to-use workflows and skills for common AppFollow tasks.

Use the AI Toolkit when you want ready-made workflows for tasks such as feedback analysis, ASO analysis, and app discovery. Use the MCP Server when you want to connect AppFollow directly to an MCP-compatible AI assistant and work with your data using natural-language prompts.

Prerequisites

  • An AppFollow account, signed in with the Google account associated with it
  • Owner or Admin permissions on your AppFollow account — required to connect the MCP Server
  • An MCP-compatible AI client (see supported clients below)

AppFollow AI Toolkit

The open-source AppFollow AI Toolkit provides ready-to-use workflows and skills built on top of the MCP Server. It's available on GitHub and used with supported AI clients such as Claude, Claude Code, and Cursor.

The toolkit helps automate common AppFollow tasks, including:

  • App discovery and tracking
  • Customer feedback analysis
  • ASO analysis
  • Keyword recommendations
  • Review trend reporting
  • Structured HTML reports and JSON outputs for further automation

Example workflows

The toolkit includes commands such as:

  • /search — Find and track apps
  • /feedback-analysis — Analyze reviews, ratings, sentiment trends, feature requests, and reported issues
  • /aso-analysis — Review keyword performance and receive ASO recommendations

These workflows produce consistent, repeatable reports that can be integrated into your existing processes.

Installing the AI Toolkit

Claude Code — GUI

  1. Go to Settings > Plugins > Add > Add from repository.
  2. Enter the repository URL: AppFollow/appfollow-ai-toolkit, then select Sync.
  3. Go to Settings > Plugins > Code > AppFollow AI Toolkit and select Install.

Claude Code — Terminal

/plugin marketplace add AppFollow/appfollow-ai-toolkit
  install appfollow-ai-toolkit

Codex

codex plugin marketplace add AppFollow/appfollow-ai-toolkit

Then inside Codex, run /plugins, select AppFollow AI Toolkit, and install it. Start a new session — the first AppFollow tool call opens the Google sign-in.

Other clients (via npx)

If you're connected through the official Claude connector, Claude Cowork, or another MCP-compatible client that doesn't have a plugin marketplace, you can still add AppFollow's skills with a terminal command:

npx skills add https://github.com/AppFollow/appfollow-ai-toolkit --yes

Then reload or restart your client so it picks up the new skills.

This requires Node.js and npx installed locally, so it's a more technical path than the plugin-based install above — most people connecting through Claude.ai or Cowork won't need it, but it's there for anyone who wants the toolkit's skills without going through Claude Code or Codex. Skills installed this way don't auto-update; re-run the command to refresh them.

AppFollow MCP Server

The AppFollow MCP Server provides secure access to your AppFollow account through Google OAuth authentication. After signing in, your AI assistant can use AppFollow tools that are available for your account without requiring API keys or custom integrations.

The MCP server is available at: https://mcp.appfollow.io/mcp.

Access is limited to the AppFollow account you authenticate with, and all requests use your existing AppFollow permissions. You must be an Owner or Admin of your AppFollow account to connect the AppFollow MCP Server.

Available tools

Depending on your AppFollow plan and permissions, the MCP Server provides access to tools for:

Category What your assistant can do
Reviews management Work with reviews and review replies
Ratings analysis Track and analyze rating trends
App rankings Check category and keyword rankings
Keywords and ASO data Review keyword performance, get ASO recommendations
Apps and collections Manage tracked apps and app groups
Reports Generate structured reports
API credits Check remaining credits via /get_credits

Your AI assistant can combine these tools to summarize customer feedback, identify trends and recurring issues, analyze rating changes, monitor ASO performance, generate reports, and answer questions using your live AppFollow data.

Once connected, try prompts like:

  • Summarize this week's new reviews for [app name]
  • What's my keyword ranking trend for [keyword] over the last 30 days?
  • Which apps in my collection had the biggest rating drop this month?

Authentication

The AppFollow MCP Server uses Google OAuth authentication.

Benefits include:

  • No API keys to manage
  • Secure access to your AppFollow account
  • Permissions scoped to your account
  • Easy sign-in from supported AI clients

Supported AI clients

The AppFollow MCP Server can be used with any MCP-compatible client, including:

Client Connect method Restart needed
Claude.ai / Claude Cowork Connectors directory No
Claude Code CLI command or plugin install Yes — new claude session
Codex CLI command or plugin install Yes — new session
ChatGPT Manual plugin entry No
Cursor mcp.json entry Yes — reload
GitHub Copilot (VS Code) .vscode/mcp.json No — click Start
GitHub Copilot CLI /mcp add No
Gemini CLI gemini mcp add command No
Gemini Enterprise Custom MCP data store (admin setup) No
Gemini app (gemini.google.com) Custom app connector

Connect to Claude

Claude.ai / Claude Cowork

The AppFollow connector is available directly in the Claude connectors directory:

  1. Open Settings in Claude.
  2. Go to Connectors.
  3. Select Browse connectors.
  4. Search for AppFollow.
  5. Select Connect.
  6. Sign in with the Google account associated with your AppFollow account.

You can also open the AppFollow connector directly: Connect AppFollow to Claude.

ChatGPT Image Aug 12, 2026, 01_00_20 PM.png
ChatGPT Image Aug 12, 2026, 01_00_30 PM.png
ChatGPT Image Aug 12, 2026, 01_00_35 PM.png

Claude Code

Run:

claude mcp add --transport http appfollow https://mcp.appfollow.io/mcp

Then authenticate by running:

/mcp

Follow the Google sign-in flow.

To install the AI Toolkit alongside the MCP connection, see Installing the AI Toolkit above.


Connect to Codex

To add the AppFollow MCP Server:

codex mcp add appfollow --url https://mcp.appfollow.io/mcp

Then authenticate:

codex mcp login appfollow

Complete the Google sign-in when prompted.


Connect to ChatGPT

To connect AppFollow in ChatGPT:

  1. Open Settings.
  2. Navigate to Security and login.
  3. Enable Developer mode.
  4. Open Plugins.
  5. Select New Plugin.
  6. Enter:
    • Name: AppFollow
    • Connection:https://mcp.appfollow.io/mcp
  7. Save the plugin and complete the Google sign-in process.

Connect to Cursor

  1. Open Cursor Settings and go to MCP (labeled Tools & MCP in newer versions).
  2. Select New MCP Server. This opens your mcp.json file.
  3. Add the following entry:

    {
      "mcpServers": {
        "appfollow": {
          "url": "https://mcp.appfollow.io/mcp"
        }
      }
    }
  4. Save the file.
  5. Cursor will open a browser window to complete the Google sign-in. Approve access to continue — Cursor stores the credentials for you.

Connect to GitHub Copilot

The AppFollow MCP Server works with GitHub Copilot in VS Code and with GitHub Copilot CLI. Both use the same remote HTTP endpoint and Google OAuth — no API key needed, and access is scoped to your own AppFollow workspace and permissions.

GitHub Copilot in VS Code

Requires VS Code 1.99 or later.

  1. Create a .vscode/mcp.json file in your repo root (or add the same entry to your user settings.json to enable it everywhere):

    {
      "servers": {
        "appfollow": {
          "url": "https://mcp.appfollow.io/mcp"
        }
      }
    }
  2. Click Start at the top of the mcp.json file.
  3. Open Copilot Chat in Agent mode and select the tools icon.
  4. The first time a tool is called, a browser window opens for the Google sign-in. After that, the AppFollow tools (reviews, ratings, rankings, ASO, keywords) appear in the tools list.

GitHub Copilot CLI

  1. Start a Copilot CLI session and run:
/mcp add
  1. Choose HTTP as the transport type.
  2. Paste the URL:
https://mcp.appfollow.io/mcp
  1. Follow the prompts to complete the Google sign-in.

Org policy. On Copilot Business or Enterprise, an admin must enable the MCP servers in Copilot policy — otherwise your MCP configuration is ignored.


Connect to Gemini

Gemini CLI
gemini mcp add --scope user appfollow https://mcp.appfollow.io/mcp --type http

Follow the Google sign-in flow when prompted.

Gemini Enterprise (Google Cloud)

AppFollow supports connecting as a custom MCP server data store in Gemini Enterprise. Setup happens on the Google Cloud side and typically requires a Cloud admin — it involves registering Gemini Enterprise as an OAuth client with your identity provider and creating a Custom MCP Server data store with the URL https://mcp.appfollow.io/mcp.

For the full setup flow, see Google's guide: Set up a custom MCP server.

Gemini app (gemini.google.com, web and mobile)

See Google's guide for the latest instructions: Connect & manage custom apps for Gemini.


Troubleshooting

Symptom Fix
Tools don't appear after connecting Fully restart the client (see the client table above) — reloading the window isn't always enough.
"Not authenticated" or sign-in loop Disconnect the connector and reconnect, making sure you sign in with the Google account tied to your AppFollow account.
"Permission denied" when calling a tool Confirm you're an Owner or Admin on the AppFollow account — other roles can't connect the MCP Server.
Tool call fails with a credits error Check remaining API credits with /get_credits — some methods are limited by plan.
Copilot ignores your MCP config On Business/Enterprise, ask your admin to enable the MCP servers in Copilot org policy.

If the issue isn't listed here, contact Support.


Feedback

We're continuously improving both the AppFollow MCP Server and the AppFollow AI Toolkit.

If you have ideas for new capabilities or encounter any issues, we'd love to hear your feedback. Your suggestions help us improve future releases and expand available functionality.

Use a built-in /submit_feedback tool to share it with our engineering team directly.

Feedback data. When you rate a session via the connector's feedback tool, you may optionally attach the session transcript and/or a log of tools used. These are sent only if you explicitly consent, and are stored by AppFollow to improve the connector. To receive a copy of your stored feedback data or have it deleted, contact help@appfollow.io with the email you used to sign in.


Methods cost and availability

For details on API credit costs per method, see Methods cost and availability.

  • Every method has its own price in API Credits
  • Method availability depends on your AppFollow subscription
  • Check your API credits consumption with /get_credits tool

Was this article helpful?

0 out of 0 found this helpful