Let your AI agent build moodboards of real websites: save, search, tag and share design references.
Scored across 9 tools
Each tool has a distinct purpose: collection creation/listing/sharing, bookmark saving/searching/updating, tag listing/suggesting, and design brief generation. The only mild proximity is between list_tags and suggest_tags, but one lists existing tags while the other proposes tags for bookmarks. No two tools appear to do the same thing.
All nine tools follow a consistent snake_case verb_noun pattern (create_collection, get_design_brief, list_collections, save_websites, etc.). Verb choices are clear and predictable across resources and actions.
Nine tools is well within the ideal 3-15 range for a bookmarking/moodboard server. Each tool covers a meaningful operation without redundant bulk.
Core workflows for saving, searching, tagging, collecting, and generating design briefs are covered, but the surface is notably one-directional. There is no delete/remove operation for bookmarks or collections, no way to unshare a collection, and collection updates are absent. These gaps could create dead ends if users need to undo or clean up saved content.
Create an empty collection (returns the existing one if the name is taken).
Inputs: name (required), description.
Turn one saved reference into design.md: palette roles, typography, layout, components, a paste-ready prompt for AI builders (v0, Cursor, Lovable, Bolt) and CSS/Tailwind tokens. Cached on the bookmark; Premium only.
Inputs: bookmark (required), refresh.
The user's collections with bookmark counts and links.
Inputs: no named parameters.
The user's tags by category (industry, style, layout…). Only these names can be applied when saving.
Inputs: no named parameters.
Save websites to the user's Moodmark library, optionally into a collection (created if missing). Moodmark captures a screenshot, title, favicon, fonts and tech stack for each; the colour palette is extracted when the user opens the library. Up to 50 URLs per call. Already-saved sites are not duplicated.
Inputs: tags, urls, items, collection, +1 more.
Find bookmarks the user already saved — by text (title, URL, note, tags, fonts, tech), colour family, tags or collection. Returns newest first.
Inputs: tags, color, limit, query, +1 more.
AI tag suggestions for up to 10 saved bookmarks, chosen from the user's own tags, each with a short reason. Suggestions only; apply the ones that fit with update_bookmarks.
Inputs: bookmarks (required).
Additive edits to saved bookmarks: add existing tags, append a note, add to a collection (created if missing). Never removes anything the user set.
Inputs: bookmarks (required), add_tags, append_note, add_to_collection.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Search real website sections from hand-picked sites by look, for design references and inspiration.
Save and organize web finds in persistent, user-controlled collections for AI assistants.
All your bookmarks in one library. Share them with your AI agents. Read-only search over your saves.
Find curated website design references, inspect motion and source, and reuse Vault components.
Add one secure layer between your agents and this server.
