Twenty-seven tools exposed via the Model Context Protocol. Drop the snippet below into Claude Desktop or Claude Code config and every tool is available to your assistant. Free tier needs no auth. Premium tier accepts USDC on Base via bearer token. The server is listed on the Official MCP Registry, Smithery, and the Glama AI catalog.
Add to your claude_desktop_config.json or Claude Code MCP config. After restart, Claude can call all 27 tools.
{
"mcpServers": {
"terminalfeed": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://terminalfeed.io/api/mcp"
]
}
}
}
Buy credits with USDC on Base, get a bearer token, pass it as a header. Same config plus an Authorization line. The full payment walkthrough is in the how-ai-agents-pay article.
{
"mcpServers": {
"terminalfeed": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"https://terminalfeed.io/api/mcp",
"--header",
"Authorization: Bearer tf_live_<your_token>"
]
}
}
}
If you are not using mcp-remote or Claude Desktop, the MCP endpoint is HTTP JSON-RPC at /api/mcp. Standard MCP discovery, tool listing, and tool calls all work over POST.
POST https://terminalfeed.io/api/mcp
Authorization: Bearer tf_live_<your_token> (optional, premium only)
Content-Type: application/json
Composed, normalized, multi-source data. Each endpoint has a deterministic credit cost.
TerminalFeed and TensorFeed are sister MCP servers with a shared credit pool. If you buy credits with USDC on Base on either site, the bearer token authenticates against both. TerminalFeed handles real-time world data (BTC, markets, earthquakes, news, prediction markets); TensorFeed handles the AI ecosystem (model pricing, status, latency probes, AI news, smart routing, cost projection). Most agents that want real-time context end up wanting both.
Quick comparison of what each server adds to your agent:
TensorFeed's MCP server is an npm package (different distribution model from this server's hosted-URL setup). Same Claude Desktop config file, just a different entry:
{
"mcpServers": {
"tensorfeed": {
"command": "npx",
"args": ["-y", "@tensorfeed/mcp-server"]
}
}
}
Premium TensorFeed tools use the TENSORFEED_TOKEN environment variable instead of an Authorization header. Same bearer token you bought on TerminalFeed works:
{
"mcpServers": {
"tensorfeed": {
"command": "npx",
"args": ["-y", "@tensorfeed/mcp-server"],
"env": {
"TENSORFEED_TOKEN": "tf_live_<your_token>"
}
}
}
}
Combined setup (both servers in one config). If you want everything, drop both entries into the same config file:
{
"mcpServers": {
"terminalfeed": {
"command": "npx",
"args": [
"-y", "mcp-remote",
"https://terminalfeed.io/api/mcp",
"--header", "Authorization: Bearer tf_live_<your_token>"
]
},
"tensorfeed": {
"command": "npx",
"args": ["-y", "@tensorfeed/mcp-server"],
"env": { "TENSORFEED_TOKEN": "tf_live_<your_token>" }
}
}
}
Source repo for TensorFeed MCP: github.com/RipperMercs/tensorfeed-mcp. Docs: tensorfeed.ai/developers. Both servers are AFTA-certified: code-enforced no-charge guarantees on errors and stale data, Ed25519-signed receipts on every paid call.
For agent-builders and tooling that wants to introspect.