Anshad Ameenza.
For machines

API & MCP endpoint

This site is built to be read and used by AI, not just people. Below are the machine-readable endpoints: a content index, the live LLM data, and an MCP-compatible endpoint so an agent can run the calculators and cite the results.

Plain endpoints

MCP endpoint

A pragmatic MCP-compatible endpoint over HTTP JSON-RPC lives at https://anshadameenza.com/api/mcp. It speaks initialize, tools/list, and tools/call, exposing:

  • llm_index — the model index (price, context, params, open vs closed).
  • vram_estimate — VRAM needed for a model size and quantization.
  • token_cost — cost of N tokens across models.
  • list_site_content — the tools, with descriptions and URLs.

A GET on the endpoint returns its descriptor. Example call:

curl -s https://anshadameenza.com/api/mcp \
  -H 'content-type: application/json' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call",
       "params":{"name":"vram_estimate",
                 "arguments":{"params_billions":744,"bits":4}}}'

Usage is open and rate-reasonable. If you build on it, a link back is appreciated.