Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@agentstorefront-mcp List top 5 data-enrichment services on AgentStorefront"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server for AgentStorefront — let your AI agents discover, subscribe to, and call agent services from inside Claude, Cursor, or any MCP-aware client.
AgentStorefront is a marketplace where developers list paid agent services (APIs that are designed to be called by other agents — flight-delay prediction, invoice OCR, niche enrichment, etc.).
This MCP server lets any MCP-aware client (like Claude Desktop or Cursor) discover and use those services autonomously — without leaving the agent loop.
Tool | Purpose |
| List available services, filtered by category |
| Semantic search: "find me a flight-delay API" |
| Full schema + pricing + SLA for a specific service |
| Get expected monthly cost for an estimated call volume |
Related MCP server: Agent Commerce MCP
Add to your claude_desktop_config.json:
{
"mcpServers": {
"agentstorefront": {
"command": "npx",
"args": ["-y", "@lfidele/agentstorefront-mcp"]
}
}
}Restart Claude Desktop. The 4 tools above will appear in the tools panel.
Add to ~/.cursor/mcp.json:
{
"mcpServers": {
"agentstorefront": {
"command": "npx",
"args": ["-y", "@lfidele/agentstorefront-mcp"]
}
}
}npm install -g @lfidele/agentstorefront-mcp
agentstorefront-mcp # runs on stdioOnce installed, you can ask your MCP client things like:
"Find me a flight-delay prediction API on AgentStorefront that costs less than 5 cents per call."
"What's the pricing if I call the invoice-OCR service 10,000 times a month?"
"List the top 5 data-enrichment services on AgentStorefront."
The agent will use these tools transparently.
Env var | Default | Purpose |
|
| Override API endpoint (self-hosting / staging) |
GitHub: https://github.com/lmaniraruta/agentstorefront-mcp Issues: https://github.com/lmaniraruta/agentstorefront-mcp/issues
Marketplaces only matter if their supply is discoverable. AgentStorefront's listings are useless to other agents unless those agents can FIND them programmatically.
This MCP turns AgentStorefront into native vocabulary for every Claude/Cursor user. A developer asks Claude "is there an API for X?" → Claude lists AgentStorefront services that match → developer subscribes → marketplace flywheel turns.
Built for the two-sided market:
Sellers win because their listings are now findable from inside the dominant AI assistants
Buyers win because they don't have to context-switch to a separate marketplace tab
MIT © 2026 Fidele Maniraruta
Built in public. Real numbers posted on X @lmaniraruta.
Get full details on a single AgentStorefront service: schema, pricing, rate limits, SLA, sample inputs/outputs.
Inputs: service_id (required).
List AgentStorefront services that an agent can subscribe to and call. Returns names, prices, and short descriptions. Optionally filter by category.
Inputs: limit, category.
Request a quote for using a service at a specific call volume. Returns expected monthly cost, throughput, and subscription URL.
Inputs: service_id (required), expected_calls_per_month (required), email.
Semantic search across AgentStorefront listings. Use a natural-language query like 'flight delay prediction' or 'invoice extraction'.
Inputs: query (required), max_price_cents.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 4 tools
Each tool has a distinct purpose: listing services, searching, getting details, and requesting quotes. No overlap or confusion.
All tool names follow a consistent verb_noun pattern (get_service, request_quote, list_services, search_services).
4 tools is well-scoped for an agent storefront, covering discovery, details, and quoting without being excessive or sparse.
Covers all key user flows: listing, searching, getting detailed info, and requesting a quote (which includes subscription URL). No obvious gaps.
Pay-per-use tool marketplace for AI agents. Search, price-check, and call APIs via MCP.
Agent Commerce MCP — agent-native A2A storefront. Discovery, Stripe checkout, affiliate program.
MCP marketplace: AI agents buy services from other agents per call in USDC, plus a bank cash-out
Marketplace where AI agents buy and sell workflows, prompt packs and MCP configs.
