Rebrand

By Rebrand

Overview

Find brand identities and visual assets in Rebrand's published design catalog. Search by creative brief, design studio, industry, visual style, typography, color or year. Explore individual assets and discover similar identities using references returned by the catalog, or attach a picture to find assets and identities that look like it. Results include gallery thumbnails and links to their sources. Designed for designers, creative teams and developers researching visual directions. Rebrand searches its published catalog; it does not edit catalog content, keep attached pictures or generate new artwork. Sign in to an existing Rebrand account with access to this feature.

Tools

search_brands

ChatGPT

Search Rebrand's published brand identities or load one identity by slug. This tool reads the catalog without modifying content. query is matched against brand write-ups and against how each identity's shots look; rankBy: visual also weighs shot captions and leans less on the write-up. Hard filters cover agency, industry, style, feel, typography, typeface, color, year, contentType (rebrand, bento or both) and editorsPick. Names and slugs are accepted; values within a filter combine with OR and different filters combine with AND. year means identity release year. Most bentos have no write-up or year, so semantic matches depend on tags and a year filter excludes entries without a year. similarTo accepts a catalog brand slug and finds similar identities, excluding the reference; with rankBy: visual it compares the look (styles, palette, shot captions) instead of the write-up, so it finds look-alikes rather than brands in the same business. image takes a picture the user attached in ChatGPT (ChatGPT supplies the file; links are not fetched) and ranks identities by how close their shots look to it, each shown at its nearest shot (matchedShot carries imageSimilarity, 0–1); filters still apply, and a query reorders those look-alikes toward its topic. Editor’s picks (editorsPick: true on a result) rank higher among identities that already match (not in unfiltered visual searches, which rank by shots), and come first when a taxonomy filter (industry, style, feel and so on) is set; identities with few or no shots rank lower; editorsPick: true keeps only the picks. Filters alone list newest first. sort: newest requires no query or similarTo. Pagination uses offset and limit. Search pages include has_more and next_offset for continuation with the same search inputs; next_offset is null when there is no next page. total is the size of the filtered set rather than a count of strong semantic matches. An empty page carries message, and relax when filters caused it: the count you would get by dropping each one. Results include taxonomy metadata, shotCount, source links and attached stills. previewUrl is the still to show: the identity’s hero shot or cover, or with rankBy: visual the best-matching shot. With a query, matchedShot is the best-matching shot, coverUrl is the identity cover, and excerpt is the matching write-up passage. A slug lookup returns the brand's markdown, metadata, cover and shot counts by asset type. Results come in rank order; the rank blends text meaning with how the stills look, so it can differ from score. score is caption or write-up similarity on one scale (roughly 6–9): under 7 is a loose association, and a note field says so when nothing fits. The attached stills are the real check. The display field describes the current client's presentation format; MCP Apps clients also receive an interactive gallery reference.

search_shots

ChatGPT

Search Rebrand's published visual assets, including symbols, wordmarks, illustrations, packaging and color palettes. This tool reads the catalog without modifying content. query is matched against shot captions (the visible subject, colors, composition and materials) and against the stills themselves. assetType is a hard catalog-type filter accepting a slug, name or group; tags is an exact tag filter. color matches a shot's tagged color family or a color named in its caption, and slug restricts results to one brand identity. similarTo accepts a shot id from a previous result and finds shots that look alike, comparing the images themselves as well as captions; those results carry imageSimilarity (0–1, how alike the pictures are, on its own scale next to score). image takes a picture the user attached in ChatGPT (ChatGPT supplies the file; links are not fetched) and finds shots whose stills look like it, nearest first with imageSimilarity; assetType, color, tags and slug still filter, and a query reorders those look-alikes toward its topic. Similarity search excludes the reference brand unless slug is set and stays in the reference asset group unless assetType overrides it; assetType: any searches every type. Filters combine with AND. Filter-only searches preserve stored order within a brand and use recency across the catalog. Pagination uses offset and limit. Search pages include has_more and next_offset for continuation with the same search inputs; next_offset is null when there is no next page. An empty page carries message, and relax when filters caused it: the count you would get by dropping each one. Results include source links and attached stills. Video assets include a still poster when available rather than the video file. Results come in rank order; the rank blends text meaning with how the stills look, so it can differ from score. score is caption or write-up similarity on one scale (roughly 6–9): under 7 is a loose association, and a note field says so when nothing fits. The attached stills are the real check. The display field describes the current client's presentation format; MCP Apps clients also receive an interactive gallery reference.