discover_servicesLABSCO SUMMARY
Three tools instead of an SDK and a pricing table.
The loop is deliberately small — find the service, call it, read the receipt — and the server keeps no cache and no local state. Money values stay decimal strings at every boundary, and an unauthenticated call gets a 401 with protected-resource metadata rather than a partial answer.
What it is
The official SynapseNetworkAI MCP server for paid API calls. It is a thin adapter over the SynapseNetwork Gateway: it does not hold custody, pricing memory or admin workflows, it just gives an agent the discover-invoke-receipt loop.
What you get
- discover_services — read-only discovery of the services available to call
- invoke_and_pay — call a paid API and settle it through the Gateway; a consequential action that should sit behind human approval in your client
- get_receipt — read the receipts recorded against the linked credential
Requirements
Either the hosted endpoint at https://mcp.synapse-network.ai/mcp for clients that speak Streamable HTTP, or the npm package @synapse-network-ai/mcp-server run locally over stdio. Local runs need SYNAPSE_AGENT_KEY, a Synapse Agent Key beginning with agt_, which the server sends as the Gateway credential header; SYNAPSE_ENV picks prod or staging and SYNAPSE_GATEWAY_URL overrides the base URL. Hosted clients authenticate with Authorization: Bearer. An OpenAI, Anthropic or xAI key is not a Synapse billing credential and must not be forwarded.
Setup effort
One command plus a key — npx -y @synapse-network-ai/mcp-server, then supply credentials
USE CASES
01
Letting an agent use an API you have not integrated
Discover the service and call it without writing SDK code or memorising routes.
02
Keeping a human in front of spending
Treat the paid call as an approval step in your client; discovery and receipts stay read-only.
03
Reconciling what an agent spent
Read the receipts recorded against the credential it used.
3 tools
Listed in project docs. Your client may show a different list.
invoke_and_payget_receiptBEFORE YOU INSTALL
What to know first
Tools you get
3 tools named in the project's documentation.
Requires
- npx on PATH
- SYNAPSE_AGENT_KEY in your environment
Where it runs
Of the 5 documented clients, 3 run the server on your machine and 2 connect to a hosted endpoint.
INSTALLATION
Connect your client
Server setup
npx
TERMINAL
npx -y @synapse-network-ai/mcp-serverHosted endpoint
ChatGPT — Remote, Claude.ai — Remote
ENDPOINT URL
https://mcp.synapse-network.ai/mcpAdd this hosted endpoint to your MCP client using the credentials listed under Requirements.
MCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"synapse-agentpay": {
"command": "npx",
"args": [
"-y",
"@synapse-network-ai/mcp-server"
],
"env": {
"SYNAPSE_AGENT_KEY": "<YOUR_SYNAPSE_AGENT_KEY>",
"SYNAPSE_ENV": "prod"
}
}
}
}Cursor
CURSOR · JSON
{
"mcpServers": {
"synapse-network-mcp-server": {
"command": "npx",
"args": [
"-y",
"@synapse-network-ai/mcp-server"
],
"env": {
"SYNAPSE_AGENT_KEY": "<YOUR_SYNAPSE_AGENT_KEY>",
"SYNAPSE_ENV": "prod"
}
}
}
}VS Code
VS CODE · JSON
{
"servers": {
"synapse-agentpay": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@synapse-network-ai/mcp-server"
],
"env": {
"SYNAPSE_AGENT_KEY": "<YOUR_SYNAPSE_AGENT_KEY>",
"SYNAPSE_ENV": "prod"
}
}
}
}
