Skip to main content
Glama

PitchSpark: Pre-Seed & Seed Deal Flow

Read one startup pitch with its score, roast and anonymised decisions

pitchspark_get_pitch_v2
Read-onlyIdempotent

Read one startup pitch by id and choose what comes back with "sections": "pitch" (the full public pitch, the same object as pitchspark_get_pitch_v1), "score" (score and roast, as pitchspark_get_score_v1), "decisions" (anonymised decision aggregates by signed-in users, as pitchspark_get_decision_summary_v1). Default: all three. Use when: an investor wants the whole picture of one shortlisted pitch in one call. Prefer the single-purpose tools when you only need one part. Do not use for: searching (pitchspark_search_pitches_v2), founder contact details, financials, decks or data rooms (never returned), investment advice. Anonymous callers can use it under the daily allowance; dimension scores and analyses need an API key, and the decisions section an investor API key. Decision data from signed-in PitchSpark users needs a PitchSpark API key from an investor account (self-declared; PitchSpark does not verify accreditation): without one every decision block is available false with reason key_required. It is anonymised: counts of distinct people only, as bands ("5-9", "10-24", "25-49", "50+"), never who; shares and rates to the nearest 0.05; the pitch's founders, team and PitchSpark staff are never counted; guest and anonymous viewers are never counted; every block is withheld (reason below_min_k) when fewer than 5 distinct people are behind it, and a split (positive vs pass, rating bands) is withheld when any part of it would be under 5; no investor comment text is ever returned. Who is counted: ratings, reading time, sections and score views count ANY signed-in non-owner user (the writers check no account type), and likes, passes and pass reasons count signed-in users who swiped or gave pass feedback (also not restricted to investor accounts); only the caller of this tool must hold an investor key. Some fields are withheld pending a review (reason hidden_pending_review, or absent from a list): the team-related pass reasons and the pass rate (the likes-vs-passes thumbs block, the search card signal, investors_considered and the card-dwell figures). Treat a withheld field as no signal. An id that does not exist, a pitch that is not public and a pitch whose founders opted out of AI agents all answer NOT_FOUND. Each pitch this tool returns counts 1 against the daily pitch allowance (the same allowance as pitchspark_search_pitches_v1), or 0 when the same API key (without a key, the same connection) already got that pitch today. Calls are logged, and each pitch returned is counted as a view.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "id": "<id>"
    },
    "name": "pitchspark_get_pitch_v2"
  }
}

Input Schema

NameRequiredDescriptionDefault
idYesThe pitch id from a search result ("id").
sectionsNoWhich parts to return. Default ["pitch","score","decisions"].

Output Schema

NameRequiredDescriptionDefault
errorNoPresent only when the call failed (isError true).
pitchNopitchspark_get_pitch_v1's pitch object (same keys), present when sections includes "pitch".
scoreNo
sectionsNo
decisionsNoAnonymised decision aggregates by signed-in PitchSpark users (not only investors); readable with investor API keys only (every block reason key_required otherwise). Counts are bands ("5-9", "10-24", "25-49", "50+"); every block is withheld below min_k distinct signed-in users (the counts cover any signed-in non-owner user, not only investor accounts).
attributionNo

Schema Changelog

Changes observed during successful MCP inspections.

  1. Added

TDQS

A4.8/5.0
Behavior5/5

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

Annotations only say read-only, idempotent, open-world; the description goes far beyond them by disclosing the key tiers required per section, the exact failure mode when no investor key is present (decisions blocks return available=false with reason key_required), the anonymisation rules (banded counts, 0.05 rounding, founders/staff/guests excluded, below_min_k withholding), review-gated fields (hidden_pending_review), NOT_FOUND conditions, and quota accounting (1 per pitch, 0 on repeat fetch, calls logged, counted as a view). This is unusually rich disclosure of side effects and constraints for a read tool.

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

Conciseness3/5

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

Purpose, triggers, exclusions, and cost are correctly front-loaded, but the anonymisation block is a dense multi-clause run-on covering counting rules, thresholds, withheld fields, and internal rationale ("the writers check no account type") in one paragraph. The content largely earns its place for a tool with this much policy, yet it could be split into short labelled chunks for far better scannability.

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

Completeness5/5

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

An output schema exists, so return shapes need not be explained; the description instead covers everything else an agent needs — permissions per section, degradation behavior, withholding reasons, quota, logging, and failure cases. Nothing material is missing for a two-parameter composite read.

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

Parameters5/5

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

Schema coverage is 100%, so the baseline is 3, but the description adds real meaning: each enum value is defined by reference to the v1 tool whose payload it mirrors, and the default of all three is restated alongside the cost implication (each returned pitch counts against the daily allowance). The id parameter is left to the schema, which already documents it.

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?

Opens with a specific verb and resource ("Read one startup pitch by id") and immediately clarifies scope via the "sections" selector, mapping each value to the equivalent single-purpose tool (pitchspark_get_pitch_v1, get_score_v1, get_decision_summary_v1). An agent can distinguish this composite read from its siblings without opening any schema.

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?

Gives an explicit positive trigger ("an investor wants the whole picture of one shortlisted pitch in one call"), an explicit routing rule ("Prefer the single-purpose tools when you only need one part"), and a negative list (no searching, contact details, financials, decks, data rooms, investment advice). That is textbook when/when-not/alternatives coverage.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources