Tensorfeed
TensorFeed.ai
TensorFeed.ai is agent fair-trade certified: open pricing, automatic no-charge on 5xx, breaker, schema fail, and stale data, Ed25519-signed receipts on every paid call, inference-only license. Built with Claude (Anthropic). Standard at /.well-known/agent-fair-trade.json.
Network of two. TensorFeed.ai and TerminalFeed.io are AFTA-federated as of 2026-04-30. A single bearer token works on both sites via the cross-Worker validate + commit rail. Two independent Ed25519 receipt keypairs, single credit ledger, one open standard. Other publishers can self-adopt by publishing a conforming manifest.
x402 + Stripe Link. TensorFeed has been x402-compliant since the agent payments rail shipped. Today we accept the
exactscheme with USDC on Base mainnet. Stripe Link (released April 2026) uses x402 withmethod=stripefor Shared Payment Tokens; supporting that method as a parallel x402 path is under evaluation. The protocol is the same; we already speak it.
Real-time AI industry intelligence for humans and AI agents. News from 15+ sources, live service status for every major LLM provider, model pricing and benchmark history, an AI agents directory, and a pay-per-call premium API settled in USDC on Base mainnet (no accounts, no API keys).
Site: https://tensorfeed.ai
Built with Claude: TensorFeed was designed by Ripper in collaboration with Claude (Anthropic). Specific systems Claude designed alongside: the agent payments rail, the active LLM probes, the GPU pricing aggregator, the OFAC sanctions screening pipeline, the routing engine, and the Agent Fair-Trade Agreement (AFTA) standard itself. We credit the collaboration openly, and the git log shows the build trail.
What this repo contains
- Web dashboard (
src/): Next.js 14 App Router, static export, deployed to Cloudflare Pages - API backend (
worker/): Cloudflare Workertensorfeed-api, attached totensorfeed.ai/api/*. Cron-driven RSS polling, status scraping, model catalog updates, daily history snapshots, daily MCP server registry telemetry - MCP server (
mcp-server/): Model Context Protocol server published as@tensorfeed/mcp-serveron npm. Exposes 19 tools (6 free, 13 paid) for use in Claude Desktop, Claude Code, Cursor, Cline, and any other MCP client. Seemcp-server/README.mdfor the full tool list and setup - Python SDK (
sdk/python/):pip install tensorfeed. Full coverage of free + premium endpoints, optional[web3]extra for one-call USDC sign-and-send - JavaScript SDK (
sdk/javascript/):npm install tensorfeed. Same surface as the Python client - Hugging Face dataset: tensorfeed/ai-ecosystem-daily. Daily JSONL snapshots of the public API committed at 08:00 UTC by
.github/workflows/huggingface-push.yml. Inference-only license per AFTA
MCP server quick start
The TensorFeed MCP server is the fastest way to plug an AI agent into the TensorFeed API.
npx -y @tensorfeed/mcp-server
Or in your Claude Desktop / Claude Code config:
{
"mcpServers": {
"tensorfeed": {
"command": "npx",
"args": ["-y", "@tensorfeed/mcp-server"]
}
}
}
Free tools work without auth. To enable the 13 paid tools (routing recommendations, news search, history series, cost projection, provider deep-dive, model comparison, MCP registry series, webhook watches, and more), add a TENSORFEED_TOKEN env var. Buy credits at tensorfeed.ai/developers/agent-payments using USDC on Base. New wallets get a 50-credit welcome bonus on first payment.
Listed in the official MCP server registry as ai.tensorfeed/mcp-server. See mcp-server/README.md for the full tool catalog, example agent queries, and pricing.
API
All endpoints live at https://tensorfeed.ai/api/*.
Free, no auth:
/api/news,/api/status,/api/models,/api/benchmarks,/api/incidents/api/agents/activity,/api/agents/news,/api/agents/status,/api/agents/pricing,/api/agents/directory/api/podcasts,/api/trending-repos/api/history,/api/history/{YYYY-MM-DD}/{type}(daily snapshots)/api/mcp/registry/snapshot(daily telemetry of the official MCP server registry)/api/health,/api/ping,/api/meta,/api/cron-status
Paid (1 credit each, USDC on Base):
/api/premium/routing,/api/premium/news/search,/api/premium/cost/projection/api/premium/history/pricing/series,/api/premium/history/benchmarks/series,/api/premium/history/status/uptime/api/premium/agents/directory,/api/premium/providers/{name},/api/premium/compare/models/api/premium/whats-new(agent morning brief),/api/premium/mcp/registry/series/api/premium/watches(webhook watches with daily/weekly digest tier)
Full documentation: tensorfeed.ai/developers and tensorfeed.ai/developers/agent-payments. Machine-readable: /llms.txt, /openapi.json, /.well-known/x402.json.
Discovery surfaces for AI agents
- Official MCP server registry:
ai.tensorfeed/mcp-server - llms.txt: https://tensorfeed.ai/llms.txt
- OpenAPI 3.1 spec: https://tensorfeed.ai/openapi.json
- x402 V2 manifest: https://tensorfeed.ai/.well-known/x402.json
- Agent-first landing: https://tensorfeed.ai/for-ai-agents
- View as Agent toggle: every page on the site exposes the underlying API JSON via a navbar toggle, so agents can introspect what powers any view
Stack
Next.js 14 (static export), Cloudflare Pages + Workers + KV, Tailwind CSS, JetBrains Mono + Inter, Resend for email alerts, Cloudflare Web Analytics. The MCP server is plain TypeScript on top of the official @modelcontextprotocol/sdk.
Development
npm install
npm run dev # Next.js dev server at localhost:3000
npm run build # Static export
npm run lint
Worker:
cd worker
npm install
npm test # 214 vitest cases, all green
wrangler deploy
MCP server:
cd mcp-server
npm install
npm run build
npm start
License
MIT. See LICENSE.
Contact
A Pizza Robot Studios project.
Server Config
{
"mcpServers": {
"tensorfeed": {
"command": "npx",
"args": [
"-y",
"@tensorfeed/mcp-server"
]
}
}
}