ListingsAPI MCP README

Listings API MCP

Connect any MCP capable AI assistant to the Listings API and manage business listings, reviews, posts and local analytics across 50 plus directories, maps and search engines.

Website · API docs · MCP page · Claude directory · Get an API key · Apify actors

What the Listings API is

Local search runs on business data. A shop, a clinic or a franchise outlet has a name, an address, a phone number, opening hours, categories and photos, and that data has to be identical on Google, Apple Maps, Bing, Facebook, Yelp, the voice assistants and dozens of smaller directories before search engines trust it enough to rank it. Keeping that data consistent by hand does not scale past a handful of locations.

The Listings API is a REST API that does this job. You store each location once, and the API syncs it to more than 50 publishers, tracks where each listing is live, finds and suppresses duplicates, pulls in the reviews those listings collect, lets you reply to them, publishes posts and offers to the connected profiles, and returns the performance data Google, Facebook and Bing report back. Agencies, multi location brands and franchise systems use it to run local presence for hundreds or thousands of locations from one place.

What this MCP server is

The Model Context Protocol lets an AI assistant call external tools. This server exposes the Listings API as a set of MCP tools, so an assistant such as Claude, Cursor, Windsurf, Gemini CLI or Cline can look up your locations, read your reviews, draft and publish replies, schedule posts and pull analytics in plain conversation, using your own account and your own data.

It is a remote server. It runs at https://listingsapi.com/mcp and uses Streamable HTTP. No local server package or build is required. Point your client at the URL and sign in through its OAuth browser flow, or use an existing API key where the client requires a header.

https://listingsapi.com/mcp

Find it in directories

The same server is listed in these directories, so you can install it from the one your client already uses:

The agent skills in this repository are listed separately; see Agent skills.

What you can do with it

Once connected, the assistant can work across five areas of your account.

Locations. Create and update location profiles, search for locations, and read their details. A location is the master record used by its publisher listings.

Listings and citations. See which publishers currently carry a location and what state each listing is in, list the sites your plan covers, pull the duplicate listings the network has detected for one location or for your whole account, and mark a listing as a duplicate so it gets suppressed, or clear that flag if the match was wrong. Voice assistant coverage is reported separately.

Reviews. Read reviews, fetch specific reviews by ID, inspect review analytics, and publish replies where the connected publisher supports them. Published replies appear on the review site.

Posts. Read existing posts and publish posts to supported connected profiles. Check the current tool contract for supported publishers and payloads.

Analytics. Read Google, Facebook and Bing performance for any location: search views, map views, direction requests, calls, website clicks and the rest, for whatever date range you ask about.

There is also a set of connected account operations for linking a Google or Facebook account, matching its profiles to your locations and creating a brand new Google Business Profile listing for a location that has none.

Destructive operations such as archiving locations and deleting posts are not exposed as MCP tools. Use the MCP guide and the server's documentation tools to check the operations currently available.

Example requests

Once the server is connected, these are the kinds of things you can ask your assistant.

• "List every location we have in Texas and tell me which ones are missing a phone number."
• "Show me all reviews under three stars from the last week across every store, grouped by location."
• "Draft a reply to the latest Google review for the Austin store, keep it under 60 words, and post it."
• "Publish an offer post about 20 percent off servicing to every location in the Northeast region, running through the end of the month."
• "Compare direction requests and phone calls for our top ten stores this month against last month."
• "Which of our locations still do not have a Google Business Profile listing?"

What you need

A ListingsAPI account. See plans and signup for production access, or check Day Pass availability below. OAuth-capable clients let you sign in without copying an API key into a configuration file.

Try it free with a day pass

Check live availability before signup; the Day Pass is offered for limited periods. While open is true, you can activate a free 24-hour sandbox at Day Pass signup. It covers up to 2 locations and syncs to 10 ListingsAPI demo directories. It never writes to real publishers. Reviews, posts, social, analytics, connected accounts, and webhooks are excluded.

Signup asks for your name, email, and company, plus acceptance of the Terms of Service and Privacy Policy. The 24 hours begin when you activate the pass through the emailed link or 6-digit code. Read the Day Pass guide for current limits and activation steps. In an OAuth client, choose Email me a sign-in link with the Day Pass email. For API-key clients, configure the Day Pass key locally using the same API header prefix. Installing a plugin or skill does not start a pass.

Keys carry an access level. A key with Read access covers every lookup and report. Creating and updating locations, replying to reviews, publishing posts and connecting accounts need a key with Write access.

The server accepts two ways of authenticating:

OAuth uses public clients, PKCE, and dynamic client registration. No pre-issued client secret is needed. Start with read access unless your task requires writes. OAuth tokens apply to the MCP endpoint; REST integrations use API keys.

Setup

Use the client instructions below. Replace API-key placeholders only in local settings, and keep credentials out of shared repositories.

Claude Code

claude mcp add --transport http listingsapi https://listingsapi.com/mcp --header "Authorization: API <your-api-key>"

Claude Desktop

Open claude.ai/directory/listings-api and click Connect. Claude runs the sign-in flow and the connector is ready in Claude Desktop and Claude.ai.

To add it by URL instead, open Settings, then Connectors, then Add custom connector, and enter https://listingsapi.com/mcp as the URL. Claude Desktop will prompt for authentication when the server first requests it.

Cursor

Install it from Cursor Directory, or add to ~/.cursor/mcp.json, or to .cursor/mcp.json inside a project:

{
  "mcpServers": {
    "listingsapi": {
      "url": "https://listingsapi.com/mcp"
    }
  }
}

Use Connect on the server entry to complete OAuth in your browser. This repository's Cursor plugin uses the same configuration. If you need API-key authentication, add "headers": { "Authorization": "API <your-api-key>" } inside the entry in local settings.

Windsurf

Open the MCP configuration file from your editor's MCP settings; the legacy Windsurf location is ~/.codeium/windsurf/mcp_config.json. Editor versions can use a different path. For a client that supports OAuth, add:

{
  "mcpServers": {
    "listingsapi": {
      "serverUrl": "https://listingsapi.com/mcp"
    }
  }
}

Complete the browser authentication prompt. For API-key authentication, add "headers": { "Authorization": "API <your-api-key>" } inside the entry in local settings. See the current Cascade MCP documentation for your editor version.

Gemini CLI

Install this repository as an extension:

gemini extensions install https://github.com/listings-api/listingsapi-mcp

Restart Gemini CLI, then run /mcp auth listingsapi and complete the browser sign-in. Use /mcp to inspect the connection and discovered tools. The extension enables OAuth and needs no API-key setting.

For an API-key setup instead of the extension, add this entry to the existing mcpServers object in ~/.gemini/settings.json:

{
  "mcpServers": {
    "listingsapi": {
      "httpUrl": "https://listingsapi.com/mcp",
      "headers": { "Authorization": "API ${LISTINGSAPI_API_KEY}" }
    }
  }
}

Provide LISTINGSAPI_API_KEY through your local environment. Use one connection configuration at a time. See Gemini's MCP documentation for authentication and configuration details.

Cline

Open MCP Servers, then Remote Servers, add https://listingsapi.com/mcp, and select Streamable HTTP. Complete Authenticate if your Cline version offers it. For manual settings, use "type": "streamableHttp"; omitting type defaults to legacy SSE. The Cline install guide includes the complete configuration, API-key fallback, and a read-only verification procedure.

VS Code with GitHub Copilot

Add to .vscode/mcp.json:

{
  "servers": {
    "listingsapi": {
      "type": "http",
      "url": "https://listingsapi.com/mcp",
      "headers": { "Authorization": "API <your-api-key>" }
    }
  }
}

Any other client

Use the streamable HTTP transport with the endpoint and header above. The server does not require SSE.

Rate limits

Requests are counted against your Listings API plan, not against this server. The Launch plan allows 10 requests per minute, Growth allows 50, and Enterprise limits are agreed per account. When you go over, the API answers with a 429 and a retry_after_seconds value that tells you how long to wait. Every error response also carries a correlation_id you can quote to support so they can find the exact request. Full detail is at listingsapi.com/docs/rate-limits.

A practical note: an assistant that is asked to "check every location" will happily issue a request per location. On the Launch plan that means a pause every ten calls, so scope the question or ask the assistant to work in batches.

Things worth knowing before you start

Connected profiles come first. Replying to a Google or Facebook review, publishing a post and reading publisher analytics all need the matching Google or Facebook profile connected to your account and matched to the location. If an analytics call comes back empty or a reply fails, an unmatched profile is the usual cause. Connect and match first, then run the workflow.

Listing creation is asynchronous. When you create a Google Business Profile listing, a successful response means the request was accepted, not that the listing is live. Google verifies and provisions on its own schedule. Check the listing status later rather than assuming it is up.

Writes are real. A reply published through this server appears publicly on the review site. A post goes live on the connected profiles. Treat write operations with the same care you would in the dashboard.

Location descriptions have a floor. A new location needs a description of at least 200 characters, which is the most common reason a create call fails on its first try.

Troubleshooting

Agent skills

Install the workflows you need from this repository with the Skills CLI. Skills supply instructions; they do not configure credentials, grant access, or run customer workflows during installation.

The five workflow skills are listed on skills.sh, and the general listingsapi skill is on ClawHub for OpenClaw.

npx skills add listings-api/listingsapi-mcp

To choose one skill directly:

npx skills add listings-api/listingsapi-mcp --skill listingsapi-integration

The integration skill retains its existing name. Its canonical source is under skills/ for standard discovery; the original listingsapi-integration/ path and ZIP are maintained as compatibility copies. Run python3 scripts/package-integration.py after editing the canonical integration skill. Read-only workflows need read access; public replies, campaign publication, profile edits, and publisher connections require approval of the concrete action and write access. Day Pass only covers sandbox locations and demo listings.

See skills publication notes for validation and release steps. GitHub availability and skills.sh indexing are separate: the skills.sh FAQ describes discovery through genuine CLI installation telemetry.

Documentation and support

• Full API reference: listingsapi.com/docs
• MCP server page: listingsapi.com/mcp
• SDKs for Python and Node: listingsapi.com/sdks
• Agent skill for integrating the API into your own code: listingsapi-integration
• Apify actors for scheduled and no code use: apify.com/listingsapi
• Support: support@listingsapi.com

About this repository

This repository holds the connection manifests that MCP directories and clients read, plus account workflow and developer integration skills. It contains no product code. The Listings API service itself is closed source and runs at listingsapi.com.

See marketplace submission notes for package paths, submission prerequisites, and current test status. Cursor and Cline setup tests have not been run for these changes.

License

The manifests in this repository are released under the MIT License. See LICENSE. Use of the Listings API itself is governed by the Listings API terms.

The conversational skill in skills/listingsapi/ is released under MIT-0 for ClawHub compatibility. The repository's root MIT license and existing integration skill are unchanged.