Discovery

MCP Server

PayWeave exposes a Model Context Protocol (MCP) server that lets AI agents discover and interact with all payment-enabled resources as native agent tools. The MCP server implements the HTTP transport and requires no authentication for discovery.

Server URLs

Scope	URL
Global (all workspaces)	/mcp
Workspace-scoped	/mcp/:workspaceId

Connecting your agent

Claude Code — run this one-liner in your terminal:

Terminal

claude mcp add payweave --transport http https://api.payweave.dev/mcp

For a workspace-scoped server:

Terminal

claude mcp add payweave --transport http https://api.payweave.dev/mcp/<workspaceId>

Cursor — open Cursor Settings → MCP → Add Server and fill in:

Plain Text

Name:      payweave
URL:       https://api.payweave.dev/mcp
Type:      HTTP

Cline — open Cline → MCP Servers → Add Server and fill in:

Plain Text

Name:      payweave
URL:       https://api.payweave.dev/mcp

You can also install the PayWeave skill which configures the MCP server automatically across any compatible agent: npx skills add payweave/payweave

Available tools

The MCP server exposes three tools:

list_resources

List all discoverable MPP-enabled resources across the platform (or within a workspace).

Parameter	Type	Required	Description
type	http \| mcp	No	Filter by resource type. Omit to return all.
limit	integer	No	Maximum results to return (default 100, max 500).

JSON

// Example response item
{
  "url": "https://api.payweave.dev/gw/gw_abc123/v1/analyse",
  "description": "Analyse text sentiment. Returns score -1.0 to 1.0.",
  "type": "http",
  "pricing": {
    "amount": "0.005",
    "currency": "USD"
  },
  "discoveryInfo": {
    "httpMethod": "POST",
    "bodySchema": { "text": { "type": "string" } },
    "exampleOutput": { "score": 0.92, "label": "positive" }
  }
}

find_resources

Search for resources using a natural language query. Results are ranked by relevance using full-text search with stemming and fuzzy matching.

Parameter	Type	Required	Description
query	string	Yes	Natural language search term (e.g. "weather forecast API").
type	http \| mcp	No	Filter by resource type.

JSON

// Agent usage example
{
  "tool": "find_resources",
  "arguments": { "query": "image classification" }
}

// Returns matching resources ranked by relevance

get_resource

Fetch full details for a single resource by its exact URL. Use this after find_resources or list_resources to retrieve the input schema and example output before making a paid call.

Parameter	Type	Required	Description
resource_url	string	Yes	The exact URL of the resource to look up.

JSON

// Agent usage example
{
  "tool": "get_resource",
  "arguments": {
    "resource_url": "https://api.payweave.dev/gw/gw_abc123/v1/analyse"
  }
}

Recommended agent workflow

Plain Text

1. find_resources  →  discover candidates by natural language
2. get_resource     →  inspect schema and example for the chosen endpoint
3. Call the endpoint URL directly with MPP payment credentials

MCP discovery manifest

GET/.well-known/mcp.json

Returns the MCP server discovery manifest listing the available server URL and supported protocol version.

LLMs.txt AI Plugin