Screen AI startups across 7 due diligence dimensions and generate a Sieve Score to evaluate investment potential. Built for VCs, angels, and founders who need structured, repeatable company assessment.
io.github.lmwharton/sieve-mcp
Local install
STDIO
1 required env var
How models use it and what it is built for.
Screen AI startups across 7 due diligence dimensions and generate a Sieve Score to evaluate investment potential. Built for VCs, angels, and founders who need structured, repeatable company assessment.
Local install — runs as a subprocess.
Configuration this server reads at startup.
Your Sieve API key (get one at https://app.sieve.arceusxventures.com/settings)
Where to find authoritative docs and source for Lmwharton Sieve.
Paste any of these into Agent Studio after connecting Lmwharton Sieve.
Common questions about connecting and running Lmwharton Sieve.
What are the 7 dimensions Sieve evaluates?
The registry describes screening across 7 dimensions for due diligence, but the specific dimension names are not detailed in the available documentation. Check the Sieve app dashboard or contact support for the full framework breakdown.
How do I get a Sieve API key to use this MCP server?
Visit https://app.sieve.arceusxventures.com/settings and generate an API key, then set it as the SIEVE_API_KEY environment variable when running the server.
Is there a cost to use Sieve for due diligence screening?
Pricing details are not included in the server documentation. Visit the Sieve app or contact Arceus Ventures for current pricing and plan options.
Can I export or share Sieve Scores with my investment team?
The MCP server retrieves Sieve Scores through the API, but export and sharing capabilities depend on your Sieve account plan. Check your dashboard for collaboration and reporting features.
What happens if my API key is invalid or expired?
The server requires a valid SIEVE_API_KEY environment variable. If authentication fails, verify your key at https://app.sieve.arceusxventures.com/settings and regenerate if needed.
MCP Playground runs 10,000+ hosted MCP servers — GitHub, Linear, Notion, Stripe, Sentry and more — across Claude, GPT, Gemini, DeepSeek and 30+ AI models. Compare model answers side-by-side, save agent presets, share runs. Zero install.
Open Agent Studio