mcp.riveterhq.com

Riveter turns web research into structured data. Give Claude a list of companies, people, or URLs and ask for new columns: product descriptions, revenue, deep analysis, company headcount, link traces, pricing, tech stack analysis, contact details, a classification, a summary, etc. Riveter runs an AI agent per row that searches, reads pages, and fills in each cell with a sourced answer.

What you can do

• Enrich rows you already have. Paste a list or point at a saved enrichment and get the new columns back. Up to 10,000 rows per run.
• Build a list from a description. "Every lawfirm in Indiana", "every YC W24 company in healthcare", "competitors of Stripe". Riveter generates the rows, and can enrich them in the same run.
• Scrape a page. Clean text or markdown from any URL, including JavaScript-rendered pages.
• Search the web. One-shot search results, or a research agent that answers a question with a structured, schema-shaped response.
• Extract records from a site. Define the fields you want and pull them from listings, directories, or catalogs as JSON.
• Monitor for changes. Run a saved enrichment daily, weekly, or monthly and get alerts or webhooks when values change.

How it works

Long runs are asynchronous. Claude starts the run, checks status, and fetches results when they are ready. Every run has an id you can come back to later, and results stay available in your Riveter account.

Setup

Click Connect and sign in to your Riveter account. Riveter creates an API key for this connection; you can revoke it at any time from Settings → API to disconnect. Runs use credits from your Riveter plan. Free accounts include trial credits.

Read-only tools (status, results, listings) are marked read-only. Tools that start runs or change saved configuration are marked as writes. Stopping a run and pausing a monitor are marked destructive.

Docs: https://docs.riveterhq.com