What you can do with Prowl
Prowl gives any AI agent 448 market intelligence tools over MCP. Here are common ways teams put it to work.
Prowl is not a dashboard you log into. It is one MCP endpoint your agent connects to — after that, the agent does the research. Below is the wiring for the two clients most people start from; the endpoint and the header are identical everywhere, so anything that speaks the Model Context Protocol works the same way.
Connect your client
- Claude MCP server — Connect Prowl as an MCP server in Claude Code or Claude Desktop and Claude gains 448 live market-data tools, billed per call from a USD wallet.
- Cursor MCP server — Add Prowl to .cursor/mcp.json and Cursor gains 448 live market-data tools over MCP, billed per call from a USD wallet. No seats.
What they have in common
Every one of these runs the same way. Your agent calls prowl_analyze for a full report, or reaches individual tools through prowl_search_tools and prowl_call_tool when it wants one specific number. There are no seats and no per-tool API keys to manage: the wallet is debited per call at provider cost plus a uniform markup, and the response carries the exact amount billed.
The endpoint is https://prowl.chat/mcp over Streamable HTTP, authenticated with an Authorization: Bearer prowl_<your_key> header. That is the whole integration surface — no SDK to install, no callback to host, nothing to redeploy when the tool catalog grows.
What your agent gets
- Live SERPs across 60+ engines — positions, features and competitors for a keyword set.
- Two independent backlink indexes (DataForSEO and Majestic), so a link claim can be cross-checked rather than trusted.
- Demand data — search volume, keyword difficulty, CPC and clickstream estimates, from two separate datasets.
- Ads and reviews — competitor creatives, review mining, listings and GMB.
- AI-surface visibility — how often a brand is mentioned across ChatGPT, Claude, Gemini and Perplexity.
What you need to start
An MCP client — Cursor, Claude Code, Claude Desktop, Codex or your own stack — and an API key. New accounts get $5 in credit, which is enough for several basic runs. The Getting Started walkthrough covers the wiring; what to expect from a report run covers cost and duration before you spend anything.