Sieve Mcp
Screen AI startups across seven due-diligence dimensions and pull a Sieve Score while researching opportunities or competitors.
Overview
Sieve MCP is a MCP server for the Idea phase that screens AI startups across seven dimensions and returns a Sieve Score via the Sieve API.
What is this MCP server?
- AI startup due diligence across 7 dimensions with a composite Sieve Score
- Sieve API integration via PyPI package sieve-mcp (stdio)
- Requires SIEVE_API_KEY from app.sieve.arceusxventures.com settings
- Version 0.1.4 — focused screening workflow for venture-style research
- Agent-callable MCP tools replace manual spreadsheet diligence tabs
- Screens companies across 7 due-diligence dimensions
- Package sieve-mcp version 0.1.4 on PyPI
- Requires SIEVE_API_KEY environment variable
Community signal: 4 GitHub stars.
What problem does it solve?
Indie builders and micro-VCs lack a fast, repeatable way to score AI startups while researching markets from the agent IDE.
Who is it for?
Builders doing competitive landscape or investment-style research on AI startups before writing specs or landing pages.
Skip if: Teams that only need generic web scrape summaries with no structured scoring model or Sieve account.
What do I get? / Deliverables
After you add Sieve MCP and an API key, your agent can run dimension-based diligence and surface a Sieve Score for named companies.
- Seven-dimension diligence breakdown for queried AI startups
- Sieve Score suitable for research memos and compare lists in the agent session
Recommended MCP Servers
Journey fit
Early journey research on which AI companies to study, partner with, or emulate belongs in idea before you commit validate/build scope. research is canonical for structured company screening and scored diligence outputs rather than shipping product code.
How it compares
Structured startup diligence MCP, not a generic CRM or code review skill.
Common Questions / FAQ
Who is Sieve MCP for?
Solo builders, analysts, and small funds using MCP agents who want scored AI startup diligence without leaving the terminal.
When should I use Sieve MCP?
Use it during idea research when comparing AI companies, validating hype, or prioritizing which startups deserve deeper validate-phase interviews.
How do I add Sieve MCP to my agent?
Install sieve-mcp from PyPI, set SIEVE_API_KEY in the MCP server env (from Sieve app settings), and register the stdio server in your MCP client config.