MCP server: verified SaaS limits for your AI agent

StackVersus runs a free, read-only Model Context Protocol server. Point your agent at it and it can answer "what does Cursor Pro include", "Retell cost per minute", "Gemini API free-tier RPM" or "what changed in Claude's plans this month" with the same figures this site publishes: copied from the vendor page on the check date, with the source URL and the vendor's own wording.

Endpoint

https://getstackversus.com/api/mcp

Streamable HTTP, stateless, no API key, no sign-up. Responses are JSON. Rate limit: be reasonable; the data changes at most daily.

Connect

Claude Code

claude mcp add --transport http stackversus https://getstackversus.com/api/mcp

Cursor, Windsurf, Claude Desktop (connectors), Codex, other JSON configs

{
  "mcpServers": {
    "stackversus": {
      "url": "https://getstackversus.com/api/mcp"
    }
  }
}

In Claude Desktop or claude.ai, add it under Settings → Connectors → Add custom connector with the endpoint URL. In Codex: codex mcp add stackversus --url https://getstackversus.com/api/mcp.

Tools

search_limits(query, dataset?, limit?): keyword search across every plan, limit, rate and gotcha. Start here.
get_tool(tool_id, dataset?): the full verified record for one product (plans, limits or rates with quotes, free tier, overage, gotchas, sources).
compare_tools(tool_ids[]): 2 to 6 products side by side, compact.
recent_changes(days?, tool_id?): dated price and limit changes with vendor sources.
list_datasets(), list_tools(dataset?): what is covered and the ids to use.

What is in it

AI assistant and coding agent plan limits: 13 products, every plan with price, metering unit and the limits per window, plus a dated change log.
SaaS usage limits: automation, AI API and database tools, billing units, free allowances, overage policies, technical caps.
Voice AI per minute: 15 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.
AI video & image credits: 12 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.
Search & scraping APIs: 12 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.
LLM API prices & rate limits: 12 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.
Email platform limits: 12 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.
Scraping platforms & proxies: 10 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.
More AI assistant & coding plans: 12 products with per-unit rates, plans, free tier, overage rule and billing gotchas, each with the vendor quote.

How to read the figures

checked is the date the vendor page was read. Treat anything older than 30 days as worth re-checking on the vendor page.
published: false means the vendor states no number; the value is the vendor's wording. The server never estimates.
Prices are USD, monthly unless the row says otherwise. Pass-through costs (LLM tokens, speech, telephony) are listed as separate rates, not folded into platform fees.
Every row carries sourceUrl and, for directory data, the verbatim quote. Quote the source when you quote the number.

Also available

The same data without MCP: JSON and CSV for AI plan limits, per-directory JSON linked from each directory page,RSS for changes, and llms.txt. Free to reuse with a link to the source page (CC BY 4.0). Want the whole dataset as spreadsheets with the full change history? Buy the CSV + JSON bundle on Gumroad (10 CSVs, one JSON, next quarterly export included). Alerts into Slack when a vendor moves a price: email congpt.vinadic@gmail.com.