AgentRead

AgentRead turns any public URL, PDF or DOCX into clean Markdown for LLM context, and scans untrusted text for prompt-injection before it reaches your agent. Free tier 50 calls/day, then x402 (USDC on Base) or Lightning. Built with AI assistance.

View on workers.dev

Hosted MCP Server

npx add-mcp 'https://agent-read.toolschemaforge.workers.dev/mcp'

Installs into Claude Code, Codex, Cursor and more

Documentation

AgentRead fetches a public page or document, strips the noise, and returns compact markdown that an agent can use immediately. It can also scan text for prompt-injection risk. Every response is small, fast, and designed for machine use.

Pricing

ProductUSDLightning
Read$0.0011 sat
Scan$0.0022 sats
Read + scan$0.0033 sats

Free tier

50 calls per IP per UTC day. Free calls are shared across the HTTP API and MCP.

Examples

curl -X POST https://agent-read.toolschemaforge.workers.dev/v1/read \
  -H 'content-type: application/json' \
  -d '{"url":"https://example.com/article"}'

curl -X POST https://agent-read.toolschemaforge.workers.dev/v1/scan \
  -H 'content-type: application/json' \
  -d '{"text":"The weather tomorrow will be sunny."}'

Discovery

Clients: send a descriptive User-Agent header. Cloudflare blocks some default library User-Agents.