Quilty

By Quilty LLC

Get New App Alerts

Open in ChatGPT

Open in Claude

Overview

Quilty is a diagnostic screenplay analysis platform. This connector lets Claude read the completed analyses you already own, so you can talk through coverage, Quilty Score, and specific scenes without leaving the chat. What Claude can do • List your completed analyses (title, year, score) • Open an analysis manifest (section index and sizes, not the full report) • Fetch one analysis section at a time • Return the Quilty Score with confidence range and pillar breakdown • Read screenplay metadata (title with year, page count, genre, content type) • Semantically search scenes in your uploaded screenplay All tools are read-only. The connector cannot upload scripts, start analyses, change settings, or spend credits. Claude only sees analyses on the Quilty account you sign in with. Scores are diagnostic assessments of the uploaded text, not predictions of box office or commercial outcome. Connect with your Quilty login (OAuth). You need a Quilty account and at least one completed analysis. Setup: https://quilty.app/docs/claude-connector

Tools

compare_analyses

ChatGPT

Compare two completed analyses the user owns. Returns both briefs plus deterministic score and pillar deltas and side-by-side strengths, issues, and actions. Does not invent narrative differences. Returns insufficient_data if either analysis is missing, unfinished, or not owned by the user.

get_analysis_brief

ChatGPT

Return one bounded overview of a completed analysis: title, diagnostic score and confidence, verdict, headline strengths and issues, priority actions, and chapter summaries with section keys. Prefer this after list_analyses. Use get_analysis_chapter or get_analysis_section to drill into evidence. Diagnostic only.

get_analysis_chapter

ChatGPT

Return one report chapter (for example storycraft, market, character) as ordered section objects. Paginated with the same 24000-character default and 60000 hard cap as get_analysis_section. Pass chapter key from get_analysis_brief or get_analysis_manifest. Follow next_cursor for the rest.

get_analysis_manifest

ChatGPT

Return the section index for one analysis: chapters, section keys, human labels, and approximate token sizes. Does NOT return section bodies — use get_analysis_section. Returns status insufficient_data if the analysis is missing, unfinished, or not owned by the user.

get_analysis_report_url

ChatGPT

Return a short-lived URL (about five minutes) for the owner's private full-analysis PDF. Use only when the user asks for the complete formatted report. Not a public share link. Returns insufficient_data if the analysis is missing or the PDF cannot be produced.

get_analysis_section

ChatGPT

Return ONE section of an analysis, paginated. Pass the section key from get_analysis_manifest. Default max_chars is 24000 (~6k tokens); hard cap 60000. Follow next_cursor for the rest. Returns insufficient_data if the section is missing. Diagnostic only — never invent missing content.

get_market_intelligence

ChatGPT

Return comparable films (title with year) and/or the IP market analysis blob. Pass include as 'comparables', 'ip', or 'comparables,ip' (default both). Diagnostic only — comparables are not a box-office forecast.

get_quilty_score

ChatGPT

Return the QuiltyScoreV5 for one analysis: score, confidence_low, confidence_high, model_agreement, pillar_breakdown, and basis. This is a DIAGNOSTIC assessment, never a prediction of commercial outcome. Returns insufficient_data if no score exists. Always surface the disclaimer field to the user.

get_screenplay_metadata

ChatGPT

Return title (with year), page count, genre, content type, and released-film match state for one screenplay. Does not return analysis section bodies.

get_talent_package

ChatGPT

Return the nested talent_summary plus user-budget and modeled-budget talent sections for one analysis, unchanged from stored JSONB. Returns insufficient_data if none of those sections exist.

list_analyses

ChatGPT

List the authenticated user's completed Quilty analyses. Returns analysis_id, screenplay_id, title with year, content_type, score, and completed_at. Diagnostic metadata only — use get_analysis_brief next for a bounded overview. Paginate with cursor.

search_analysis

ChatGPT

Case-insensitive keyword or phrase search across the completed analysis report (not the screenplay text). Returns section keys, labels, match counts, and short snippets. Use search_screenplay to search scene text. Diagnostic only — do not invent matches.

search_screenplay

ChatGPT

Semantic search over the SCREENPLAY TEXT (scenes), not the analysis report. Returns ranked scene snippets with heading and similarity. Prefer many small targeted searches. Returns insufficient_data if the screenplay has no extractable scenes.