MCP tool
archive_report_brief
One-call complete archive research package for a document, PDF, or editorial brief.
The archive_report_brief tool is provided by Colour Memory. Install that server in your AI client and the assistant can call it directly.
Servers exposing archive_report_brief
From the publisher:One-call complete archive research package for a document, PDF, or editorial brief. Input: title, audience, themes, archives to draw from, things to avoid, number of colours. Output: ranked colour cards with full provenance, story order, source confidence flags, pull quote, CTA line, CSS tokens, image prompt for Midjourney/Flux/DALLE, editorial argument, weakest and strongest entries identified. R
Install Colour Memory →Parameter Type Description titlestring Document title e.g. 'The Colours of Georgian Power' audiencestring Target audience e.g. 'serious Georgian collector' themes*array Research themes e.g. ['racing silks', 'EIC trade', 'Keats'] archivesarray Archives to search e.g. ['RacingSilks', 'EIC', 'Keats', 'Dickens'] avoidarray Topics to suppress e.g. ['arsenic wallpaper', 'Wedgwood blue'] n_coloursinteger Number of colours to return (default 8, max 16) strict_sourcesboolean Only return entries with named primary sources (default true) minimum_claim_strengthstring Hard evidence-grade floor: A, B, C, D, or E. Never silently relaxed. confidence_thresholdnumber Minimum confidence score (0.0-1.0) for an entry to be included. Default 0.5 filters out weakly-matched stubs. period_startinteger Hard period gate: exclude entries dated before this year. Default 0 (no lower bound). period_endinteger Hard period gate: exclude entries dated after this year. Default 9999 (no upper bound). allow_partialboolean If false (default), returning fewer cards than requested sets ok=false with error_code INSUFFICIENT_PALETTE_RESULTS instead of a silent short 'success'. Set true to accept a short result. family_card_modestring How near-identical cultivars/records are presented. 'auto' (default): current behaviour, groups into family cards above the standard ΔE threshold. 'always': widens the clustering threshold so more nea max_family_card_sharenumber Conservative mode only: maximum fraction (0.0-1.0, default 0.5) of presented cards that may be family cards before excess families are suppressed to hero cards. Ignored in auto/always modes. output_formatstring Reserved for future use -- currently accepted but has no effect on the response shape (the endpoint always returns the full report). strict_semantic_colourboolean Default false. When true, excludes entries whose quoted colour language contradicts their stored hex (e.g. a record naming 'Prussian blue' whose hex is actually a pale cream) rather than presenting th
Related tools
The same action under a namespaced name on other servers.
Tool lists are read from the live servers (or, for packaged servers, from their documentation) each morning, so this reflects what they exposed at the last refresh. Search all tools.