Skip to main content
Glama
lmwharton

lmwharton/sieve-mcp

sieve_screen

Analyze a startup across 7 dimensions (Innovators, Market, Product, Advantage, Commerce, Traction, X-Factor) to get a quick screen result. Use with data room documents (v3) or directly with company name/website (v2). Returns an analysis ID.

Instructions

Run a Sieve IMPACT-X Quick Screen on a startup.

Analyzes the company across 7 dimensions (Innovators, Market, Product, Advantage, Commerce, Traction, X-Factor) and returns an analysis ID. Takes 2-5 minutes to complete. Upserts -- if the company was previously screened, returns the existing deal (set confirm=true to re-screen).

Two ways to use:

  • v3 (recommended): First add documents with sieve_dataroom_add, then call sieve_screen(deal_id=...) to analyze everything in the data room.

  • v2 (legacy): Call sieve_screen(company_name=..., website_url=...) directly. At least one of website_url or pitch_deck_text is required in this mode.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
company_nameNoName of the startup to screen (v2 flow, or to create new deal).
deal_idNoScreen an existing deal by ID (v3 flow -- use after sieve_dataroom_add).
website_urlNoCompany website URL (v2 flow).
pitch_deck_textNoExtracted pitch deck text (v2 flow).
descriptionNoBrief company description (optional).
confirmNoSet to true to re-screen an existing deal.

Output Schema

TableJSON Schema
NameRequiredDescriptionDefault

No arguments

Implementation Reference

  • The MCP tool handler function for sieve_screen. It is registered as a FastMCP tool, accepts parameters like company_name, deal_id, website_url, pitch_deck_text, description, and confirm, and delegates to client.screen().
    async def sieve_screen(
        company_name: str = "",
        deal_id: str = "",
        website_url: str = "",
        pitch_deck_text: str = "",
        description: str = "",
        confirm: bool = False,
    ) -> dict:
        """Run a Sieve IMPACT-X Quick Screen on a startup.
    
        Analyzes the company across 7 dimensions (Innovators, Market, Product,
        Advantage, Commerce, Traction, X-Factor) and returns an analysis ID.
        Takes 2-5 minutes to complete. Upserts -- if the company was previously
        screened, returns the existing deal (set confirm=true to re-screen).
    
        Two ways to use:
        - v3 (recommended): First add documents with sieve_dataroom_add, then
          call sieve_screen(deal_id=...) to analyze everything in the data room.
        - v2 (legacy): Call sieve_screen(company_name=..., website_url=...) directly.
          At least one of website_url or pitch_deck_text is required in this mode.
    
        Args:
            company_name: Name of the startup to screen (v2 flow, or to create new deal).
            deal_id: Screen an existing deal by ID (v3 flow -- use after sieve_dataroom_add).
            website_url: Company website URL (v2 flow).
            pitch_deck_text: Extracted pitch deck text (v2 flow).
            description: Brief company description (optional).
            confirm: Set to true to re-screen an existing deal.
        """
        return await client.screen(
            company_name=company_name,
            deal_id=deal_id,
            website_url=website_url,
            pitch_deck_text=pitch_deck_text,
            description=description,
            confirm=confirm,
        )
  • The @mcp.tool decorator that registers sieve_screen as a FastMCP tool with annotations.
    @mcp.tool(
        annotations={
            "readOnlyHint": False,
            "destructiveHint": False,
            "openWorldHint": True,
        }
    )
  • The client.screen() helper function that builds the request body and calls the Sieve API POST /screen endpoint with a 30-second timeout.
    async def screen(
        company_name: str = "",
        deal_id: str = "",
        website_url: str = "",
        pitch_deck_text: str = "",
        description: str = "",
        confirm: bool = False,
    ) -> dict[str, Any]:
        """Start a Quick Screen analysis (upserts — rescreens if deal exists)."""
        body: dict[str, Any] = {}
        if company_name:
            body["company_name"] = company_name
        if deal_id:
            body["deal_id"] = deal_id
        if website_url:
            body["website_url"] = website_url
        if pitch_deck_text:
            body["pitch_deck_text"] = pitch_deck_text
        if description:
            body["description"] = description
        if confirm:
            body["confirm"] = True
        return await _request("POST", "/screen", json_body=body, timeout=30.0)
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations indicate readOnlyHint=false, destructiveHint=false, and openWorldHint=true. The description adds behavioral context: the tool takes 2-5 minutes to complete, and it upserts (returns existing deal if previously screened unless confirm=true). This goes beyond the annotations, though it does not detail the exact impacts or error states. No contradiction found.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is concise yet comprehensive, with a clear structure: introduction, timing and upsert behavior, then two usage modes. Every sentence adds value, and the formatting (bullets, indentation) aids readability. No unnecessary words or redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

Given the tool's complexity (two modes, 6 parameters, output schema exists), the description is largely complete. It covers workflows, parameter requirements, timing, and upsert behavior. The output schema is not provided, but its existence is noted, so the description need not detail return values. A slight shortcoming is not mentioning what happens if both flows are used simultaneously (though schema might enforce), but overall it is sufficient.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3. The description adds meaning by explaining how parameters relate to the two workflows (e.g., deal_id for v3, company_name/website_url for v2) and specifies the requirement that at least one of website_url or pitch_deck_text is needed in v2. This contextualizes parameters beyond the schema's individual descriptions.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description clearly states the tool's purpose: 'Run a Sieve IMPACT-X Quick Screen on a startup.' It specifies the 7 dimensions analyzed and that it returns an analysis ID. It also distinguishes two distinct workflows (v3 and v2), making the purpose and resource well-defined.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description explicitly provides usage guidelines: when to use v3 (after sieve_dataroom_add) vs v2 (directly with company_name/website_url), and notes that at least one of website_url or pitch_deck_text is required in v2. It also explains the behavior of the 'confirm' parameter for re-screening. This gives clear when-to-use guidance and differentiates from sibling tools.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Install Server

Other Tools

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/lmwharton/sieve-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server