Find new ways to wear what you already own with Trove, your personal AI stylist.
This listing identifies a third-party MCP server. MCPBundles is not affiliated with, sponsored by, or endorsed by the service or its operator unless stated on this page.
This server may ask you to sign in when your AI client connects.
Last probed Nov 3, 2026 · mcp.trovefinds.my
21tools discovered
For AI agents — MCPBundles Directory API
For MCP client apps (human connect):
MCP connect URL: https://mcp.trovefinds.my/mcp
Skill page: https://www.mcpbundles.com/skills/trovefinds-mcp-695ca1059e
Explore capabilities (21)
read 11
write 10
Refresh
4
Get
3
Prepare
2
Show
2
Other capabilities
10
Get Trove context
read
get_trove_context
Read the connected user's profile, explicit Fashion Memory, wardrobe counts, and pending wardrobe ingestions. Use browse_wardrobe when the user wants to see the actual garments.
Get Trove context — Read the connected user's profile, explicit Fashion Memory, wardrobe counts, and pending wardrobe ingestions. Use browse_wardrobe when the user wants to see the actual garments.
Browse Trove wardrobe — Use once whenever the user asks how many garments they have, what is in their wardrobe, wants to browse a category, or asks whether a just-imported garment is visible. Return the count and show the real wardrobe garments in the inline Trove gallery; after a successful result, do not call this tool again for the same request. A follow-up asking to style these garments uses prepare_wardrobe_styling, not native image generation.
Refresh Trove wardrobe gallery — UI-only read that refreshes an existing wardrobe gallery at its current category and offset with fresh image URLs.
Paginate Trove wardrobe gallery — UI-only read used by the wardrobe gallery's previous and next page controls. It replaces the current gallery contents without creating another model-visible UI.
Submit wardrobe item — Add one garment from an uploaded JPEG, PNG, or WebP. Inspect the complete image, identify the requested garment, and supply its normalized bbox (x, y, w, h) plus garment analysis. Trove pads the box and crops before cutout. Include all visible garment edges; use separate calls/boxes for different garments in one photo. Explicitly set worn_by_person for human or mannequin body parts within or immediately adjoining the target box (Trove adds 5% padding), including feet inside shoes even without a
Upload try-on photo — Privately upload a clear full-body JPEG, PNG, or WebP photo and make it the current photo for future try-ons.
Get wardrobe ingestion — Read ingestion progress and availableInPlugin. Once availableInPlugin is true, the garment can be browsed and included in outfits using its crop while cutout continues. Stop waiting for cutout before styling; do not claim App import or cutout is complete. A later wardrobe/Look refresh can show the finished cutout. A request to style this garment continues with prepare_wardrobe_styling even without a new @mention, not native image generation; missing wardrobe pieces must be reported, not invented
Get your MCP into directories
A working endpoint is step one. Directory coverage is the coordinated launch across ChatGPT, Claude, Cursor, the MCP Registry, and community indexes.
Find new ways to wear what you already own with Trove, your personal AI stylist.
How do I connect Trove to my AI agent?
Use the MCP endpoint listed on this page in your MCP client configuration. One-click install pills support Claude, Cursor, VS Code, and other hosts. Copy the remote MCP URL if your client needs a manual entry.
How many tools does Trove provide?
Operate this server?
Request removal if you do not want Trove listed here, or claim the listing if you want to manage it on MCPBundles.
Operate Trove?
This server appears in the MCPBundles directory. Verify you operate it to take over the listing — name, description, logo, contact email, and skill content. We email a 6-digit code to a maintainer address your server publishes in /.well-known/security.txt or /.well-known/mcpbundles.json. Free, takes about a minute.
MCPBundles publishes this directory for MCP discovery. Except where we host or operate an offering, third-party services run under their own terms. Product and company names on this page are used in a descriptive, identifying way (including under nominative fair use where applicable); they remain the property of their owners. Nothing here grants you rights in those marks, and nothing here is an offer to sell a third party's services. Terms · Request removal
Prepare wardrobe styling — Use for Trove wardrobe outfit requests and revisions, including plain follow-ups about a just-imported garment or previous Look without another @mention. Styling an outfit is not a request for a generated outfit/person image: do not use native image generation or try-on tools. Return up to 40 varied real wardrobe candidates, contact sheets, recent garment combinations and explicit preferences. Read all sheets and recentLooks; compose only returned aliases with meaningful differences in core garm
See more wardrobe candidates — After inspecting the current styling contact sheets, request up to 12 unseen garments in one category if suitable choices are missing. At most two expansions per candidate set, 64 total candidates. Existing aliases stay stable. Inspect every new sheet; this does not create a user-facing gallery. Reuse requestKey only for a transport retry with the same category.
Validate and persist outfits — Data-only validation and persistence after composing or revising Looks using current candidate aliases; this tool never displays a widget. Trove generates outfitSummary and stylingNote from the actual saved garments. Do not submit garment prose in stylingNote (ignored); optional wearingNotes bind supported actions to selected aliases. Revisions require lookId and expectedRevision and the complete desired wearingNotes list. Exact recent or same-batch duplicate combinations return DUPLICATE_LOOK_C
Show Trove Looks — After validate_and_persist_outfits succeeds, call exactly once with all returned lookIds in one array to display one inline Look gallery. This read-only tool reads existing owner-scoped Looks and never creates or revises outfits, generates images, or consumes try-on quota. Do not call it for failed proposals or repeat it to inspect, verify, or refresh an already displayed gallery. The host displays the real garment images automatically; keep IDs internal. An explicit try-on request continues sep
Refresh Trove Look gallery — UI-only read that reloads the displayed durable Looks with fresh image URLs. Does not generate boards, try-ons, or new Looks.
Get saved Looks — Use once when the user wants to see their saved outfits or favorite Looks. Read shared mobile and Plugin favorites and show them directly in the inline Trove gallery. Keep the returned exact lookId available for a later request such as trying on a named saved Look; do not report a connection failure just because a title needs matching. Do not generate new outfits, boards, or try-ons to display favorites.
Refresh saved Trove Looks — UI-only read that reloads the current saved Looks and renews their image URLs inside the existing gallery.
Save Look — Idempotently save a durable Agent or mobile Look to shared favorites.
Unsave Look — Idempotently remove a Look from shared favorites.
Record styling feedback — Record behavior and update Fashion Memory only from explicit structured signals. Use a neutral dimension key, one attribute value, and positive (prefer) or negative (avoid) direction. For a reversal, send the same dimension/value with the new direction; never invent like_color/avoid_color keys or negate an avoidance key. Temporary experiments and ordinary praise do not update memory.
Prepare try-on — Only for an explicit user try-on request in Work mode, never an ordinary wardrobe styling request. Reference preparation only: return exact short-lived user and garment images, deterministic preparation state, content mapping, and a strict fidelity manifest for one selected Look. This MCP tool does not generate a final image. In ChatGPT Work mode, use the native image generation/editing tool only after ready_for_host_edit with every returned reference, including footwear; in ordinary Chat mode,
Generate try-on — Ordinary Chat mode try-on generation: call exactly once for an explicit try-on request with the selected Look. Trove's backend generates or reuses the private cached image and returns tryonResultId. This is a data-only tool: it does not display a widget or return an image to inspect. When status is ready, call show_tryon_result exactly once with the returned lookId and tryonResultId, then stop. Do not call generate_tryon again to inspect, verify, display, or repair a successful result. On tryon_
Show try-on result — Display one saved backend try-on image in the built-in Trove Try-on Result widget. Call exactly once after generate_tryon returns status ready; copy its lookId and tryonResultId exactly. This read-only tool never generates or edits an image. On success the host displays the image automatically and the user's request is complete: stop, do not call further tools to inspect, verify, redisplay, or repair it, and do not embed a separate image or link. Image bytes are app-only metadata and are not ava
Reload saved Trove try-on image — UI-only read that renews the image link in an existing try-on widget. Reads the exact saved result after owner, current photo, Look, and privacy checks. Never generates an image or creates another widget.
MCPBundles probed 21 tools on the live server. The tool list on this page reflects what was discovered at the last refresh — connect your client to see the full set available to your session.
What authentication does Trove require?
Trove may require signing in to the provider before tools can run. Connect through MCPBundles or your MCP client and complete any provider login when prompted.
Don't want this listed?
Request removal from the MCPBundles directory. We stop publishing this listing on mcpbundles.com.
MCPBundles is an independent platform built on the open Model Context Protocol standard. MCPBundles is not affiliated with, sponsored by, or endorsed by Anthropic PBC.