// what is mcp

MCP, explained.

MCP, the Model Context Protocol, is an open standard for connecting AI agents to tools. One endpoint, one schema, and your agent can call a tool directly instead of you wiring up a custom integration. This page explains what MCP is, and how Tulimoa's own MCP server works.

Add mcp.tulimoa.com Browse the directory

Open MCP standard
EU hosting, Frankfurt
Read access, no account

// how mcp works

One endpoint. A shared schema. Direct tool calls.

MCP standardises the wiring between an AI agent and the tools it uses. Instead of a bespoke integration per tool, the agent speaks one protocol to one endpoint.

One endpoint

A tool provider runs an MCP server at a single URL. Your agent connects to that URL and needs nothing else to start using it.

A shared schema

The server describes its tools in a common format: what each tool does and which parameters it takes. Your agent discovers this list automatically, so it knows what it can call.

Direct tool calls

The agent calls a tool with typed parameters and gets a structured result back. No screen scraping, no glue code, no guessing at an API by hand.

// tulimoa's mcp server

A live MCP server for the Tulimoa directory.

Tulimoa runs its own MCP server. Point your agent at it and it can search the curated directory of agent-ready software from inside your agent. Read access is live today and needs no account.

Endpointmcp.tulimoa.com/mcp

One URL over Streamable HTTP. Add it once, and your agent can search, open, and list directory entries on demand.

AccessRead is authless

The three read tools work without any login, because the directory is public. The two write tools need an OAuth sign-in with a write scope, bound to your Tulimoa account.

HostingEU, Frankfurt

Built in Germany, hosted in the EU. The read tools serve public catalog data, gated so only approved, published listings are returned.

// the tools

Five tools over one endpoint.

Three read tools work without an account. Two write tools let you submit or edit your own listing after an OAuth sign-in. Every write goes back to review before it is public.

search_listingsno account

Find tools by free text and filters. Returns approved, published listings.

querycategorypricing_modelmcpeu_onlysortlimit

get_listingno account

Full detail for one tool by its slug.

slug

list_categoriesno account

The category ids and labels used by the filters.

no parameters

submit_listingwrite login

Submit an AI or MCP tool on behalf of the logged in user. Needs a logo URL and a short description of how it uses AI. Choose how it goes live: free with the Tulimoa badge (returns the badge code) or the $5 launch package with the user's consent (returns a checkout link). Already listed websites are rejected. Public after review, up to five per day.

nameurlshort_descriptioncountrycategorymcppricing_modellogo_urlai_usagevisibilitywithdrawal_consent

edit_listingwrite login

Update the fields of a listing you own. Any edit sends it back to review before it is public again.

slug

// connect it

Add Tulimoa's MCP server in one line.

Pick your client, paste the line, and your agent can search the directory right away. Read access needs no account.

claude code

claude mcp add --transport http tulimoa https://mcp.tulimoa.com/mcp

openclaw

openclaw mcp add tulimoa --url https://mcp.tulimoa.com/mcp --transport streamable-http

mcp.json

{
  "servers": {
    "tulimoa": { "type": "http", "url": "https://mcp.tulimoa.com/mcp" }
  }
}

Endpoint: https://mcp.tulimoa.com/mcp · Streamable HTTP · read access is authless.

// two servers

This is the directory server. Tulimoa Memory is the other one.

The page you are reading is about Tulimoa's MCP server, which searches the directory. Tulimoa Memory is a separate MCP server at memory.tulimoa.com/mcp that gives your agent a persistent memory. It stores experiences by day, consolidates them into knowledge once a night, and recalls them by meaning and keyword. It runs with a key and costs $0.002 per call.

Read the Memory docs

// questions

Straight answers.

What is MCP?

The Model Context Protocol is an open standard for connecting AI agents to tools. One endpoint, one schema. A tool provider runs an MCP server, and any MCP-capable agent can call its tools without a custom integration.

What is an MCP server?

It is the tool side of MCP: a server that exposes a set of tools over the protocol. Tulimoa's MCP server exposes the directory, so an agent can search and read listings. Other companies run their own servers for their own products.

Do I need an account?

Not for reading. The three read tools are authless because the directory is public. Submitting or editing a listing needs an OAuth sign-in with a write scope, bound to your Tulimoa account.

How is this different from Tulimoa Memory?

This MCP server searches the Tulimoa directory, and reading needs no account. Tulimoa Memory is a separate MCP server that gives your agent a persistent memory across sessions, used with a key and paid per call. Different jobs, and you can add both.

Where is it hosted?

In the EU, in Frankfurt. Tulimoa is built in Germany. The read tools serve public catalog data, gated so only approved, published listings come back.

What does it cost?

Reading the directory over MCP is free, and so is browsing it on the web. There is nothing to pay to add the server to your agent.

Add Tulimoa's MCP server to your agent.

One line, no account for read access, and your agent can search a curated directory of agent-ready software.

Add mcp.tulimoa.com Browse the directory