Build with AI agents
Connect Claude Code, Cursor, or ChatGPT to the hosted MCP server and let them read the docs, call the API, and write the integration for you, scoped by your API key.
Hosted MCP
MCP Server
Add secure, consent-driven access to your API with Listings API's hosted MCP server. Connect Claude Code, Cursor, or ChatGPT in one command.
Connect your client
The server is hosted, so setup is configuration only. Pick your client and replace the placeholder with your API key. In a client that supports OAuth 2.0 you can skip the key and sign in instead.
One command in your terminal. Nothing to install or host.
claude mcp add --transport http listingsapi https://listingsapi.com/mcp --header "Authorization: API <your-api-key>"Add to ~/.cursor/mcp.json.
{
"mcpServers": {
"listingsapi": {
"url": "https://listingsapi.com/mcp",
"headers": { "Authorization": "API <your-api-key>" }
}
}
}Add to ~/.codeium/windsurf/mcp_config.json.
{
"mcpServers": {
"listingsapi": {
"serverUrl": "https://listingsapi.com/mcp",
"headers": { "Authorization": "API <your-api-key>" }
}
}
}Any client that speaks stdio bridges through mcp-remote.
{
"mcpServers": {
"listingsapi": {
"command": "npx",
"args": [
"mcp-remote",
"https://listingsapi.com/mcp",
"--header",
"Authorization: API <your-api-key>"
]
}
}
}Resources for agents, all live today
listingsapi.com/mcp The hosted MCP server itself, streamable HTTP, reached with an OAuth token or an API key. llms.txt Markdown index of the developer docs for agents; llms-full.txt has the complete reference in one file. /auth.md The agent auth guide: who registers, where keys come from, how OAuth and the header work. SKILL.md A loadable skill covering the core endpoints, auth, and verification flow. openapi.yaml Every endpoint, parameter, and response shape. The source of truth agents code against. /.well-known/api-catalog RFC 9727 discovery document linking the API, its spec, and its docs. Two ways to authenticate
Both give the agent an account scoped credential. Pick whichever your client supports.
Sign in with OAuth 2.0
Add the server in a client that supports OAuth, such as Claude. You get a sign in page and a consent screen listing what the agent may read and change. Approving it creates one API key named after the client, and you can see it in your dashboard.
Authorization: Bearer <access token> Or pass an API key
Create a key in the dashboard and put it in the header your client sends. Use this when the client has no OAuth support, or when the agent runs on a server with no person at the keyboard.
Authorization: API <key> Before you start
You set up once, then the agent takes it from there.
Pick how the agent signs in
OAuth 2.0 if your client supports it, so you approve access on a consent screen. Otherwise create an API key in the dashboard, one per agent, so each can be revoked on its own.
listingsapi.com/signup Hand the agent its context
Give it the base URL https://listingsapi.com and pointers to llms.txt and auth.md so it can orient itself. With OAuth the client holds the token for you, so there is no key to pass around.
Authorization: Bearer <token> or API <key> Ask for the first location
The agent reads the docs, makes the calls, and verifies its own work through the sync status endpoint.
locations · reviews · posts · analytics Ready to use prompts
Paste one into your connected agent and swap in your details.
Publish a business
Create a location for <business name> at <address> using the Listings API MCP server, then poll the premium listings status until Google Business Profile reports SYNCED and give me the live listing URL.
Work the review queue
Pull the latest reviews for location <id> through the Listings API MCP server, draft a reply for each review under 4 stars, show me the drafts for approval, and post the approved ones.
Build a report
Read last month of Google and Bing analytics for every location through the Listings API MCP server and summarize views, searches, and actions per location in a table.
FAQs
What is the Listings API MCP server?
The MCP server gives AI agents direct access to the Listings API. Connect it once and an assistant like Claude Code, Cursor, or ChatGPT can read the docs and call the API directly, scoped to exactly what your key allows.
How do I connect it?
Point your agent at https://listingsapi.com/mcp. There is no local process to run and nothing to host yourself. Clients that support OAuth 2.0 will send you to a sign in page and a consent screen. Everything else uses an Authorization header with an API key.
How does authentication work?
Two ways, and both end up at the same place. With OAuth 2.0 you sign in, read what the agent is asking for, and press Allow. That creates one API key named after the client and hands the client a short lived access token, which it refreshes on its own. With the header method you create the key yourself and paste it into your client config. Either way the server sees a key that belongs to your account.
What can an agent do through it?
Read locations, listing status, reviews, posts, and analytics. Write access adds creating and updating locations, replying to reviews, and publishing posts. It cannot delete locations, remove posts, or disconnect your Google and Facebook profiles, because those tools are not on the server at all. The agent also fetches the OpenAPI spec and llms.txt so it can write correct integration code.
Is it secure?
The agent gets exactly what you granted and nothing more. Read access cannot change anything. Every connection shows up as its own API key in your dashboard, so you can see what is connected and delete any one of them to cut it off straight away.
Let an agent write the integration
Connect the MCP server, hand over a scoped key, and ask for your first location today.