Ocr MCP server
{
"mcpServers": {
"pennyocr": {
"type": "http",
"url": "https://mcp.pennyocr.com"
}
}
}{
"mcpServers": {
"pennyocr": {
"type": "http",
"url": "https://mcp.pennyocr.com"
}
}
}{
"mcpServers": {
"pennyocr": {
"type": "http",
"url": "https://mcp.pennyocr.com"
}
}
}{
"servers": {
"pennyocr": {
"type": "http",
"url": "https://mcp.pennyocr.com"
}
}
}{
"mcpServers": {
"pennyocr": {
"type": "http",
"url": "https://mcp.pennyocr.com"
}
}
}file .mcp.json (in your project root)file claude_desktop_config.jsonfile ~/.cursor/mcp.jsonfile .vscode/mcp.jsonfile ~/.codeium/windsurf/mcp_config.jsonversion 0.1.0
What it does
An MCP server that extracts text and markdown from PDFs and images via URL, with exact cost tracking and hard spend limits. Designed for AI agents that need to read documents while maintaining budget control.
- Extract structured text from multi-page PDFs, preserving table formatting as markdown.
- Read images by URL and convert them to plain text or markdown without local file uploads.
- Preview extraction costs before processing, then set spend caps to prevent budget overruns in autonomous workflows.
Tools this server exposes
| Tool | Does |
|---|---|
| read_document | Extract content from a PDF or image URL, returning markdown or text with page count and exact cost. |
| estimate_cost | Check page count and exact cost for a document without processing it (free). |
| check_balance | View remaining credits and free-tier allowance (free). |
Before you enable it
An MCP server runs with whatever access you give it. Start with a development account rather than production, and check which tools the server exposes before you let an agent call them unattended.