Archive Creator Search Is Still Live—But AI Only Ranks the Candidates
Archive’s Creator Search is not a new 2026 release: the company introduced it in September 2024, and it remains available alongside the Super Search UGC tool. The important update is a clearer account of what the AI does and does not decide: Archive’s current Creator Search documentation lists the feature on its Startup, Growth, Enterprise, Agency and Custom plans, but defines it primarily as a visual ranking system supported by separate filters and vetting data.
Super Search likewise remains active, but it searches content already held in an Archive library rather than the wider social web. Together, the tools can reduce a large discovery task to a workable set of creators or assets; they do not establish that a creator is commercially suitable, that an asset may legally be reused or that a result set is complete.
The 2024 release addressed discovery, not final selection
Creator Search entered Archive’s product lineup on September 16, 2024. In the original Creator Search announcement, Archive presented two ways to begin: marketers could describe the creator they wanted or upload an image to find accounts publishing visually similar material.
That basic proposition has survived. The current product covers Instagram and TikTok profiles, combines content queries with filters such as location, follower count, engagement, verification status and contact availability, and lets teams place potential partners in vetting lists. It turns open-ended browsing into a more structured candidate pool.
A place near the top of that pool is not an endorsement. Visual relevance does not establish audience authenticity, contractual availability, brand safety, usage rights or probable campaign performance. Each of those questions requires evidence beyond the similarity ranking.
Visual resemblance and commercial fit remain separate tests
The system evaluates content themes and formats as well as presentation style and aesthetics. Follower totals, engagement ranges and other filters can narrow the pool, but they are distinct from the visual signals used to determine similarity. A creator can therefore look highly relevant to a prompt without meeting the campaign’s business requirements.
The “Search similar creator” function also has a narrower basis than its name might imply. It selects one photo or video associated with a creator and uses that asset to locate accounts with comparable content; it does not evaluate the creator’s entire catalogue as a single body of work. Reviewing several recent posts is still necessary before treating one matched asset as representative.
The searchable index now includes profiles with at least 1,000 followers or subscribers. That threshold is lower than the 2,000-follower floor stated in Archive’s earlier public positioning, but it does not mean every small account known to other parts of the platform will appear in Creator Search. Social-listening and creator-management datasets can contain profiles that the discovery index excludes.
Ranking behavior can change what marketers think they found
Natural-language search sorts candidates by relevance rather than enforcing every phrase as a strict pass-or-fail requirement. Less relevant profiles can remain farther down the results, so the total returned should not be interpreted as the number of creators satisfying every element of the request.
Observable creative requirements and measurable constraints therefore play different roles. A hypothetical brief might look for creators who regularly demonstrate compact kitchen products, while separate filters restrict the pool by country, audience size and engagement. The prompt identifies content and presentation patterns; the filters impose conditions that visual similarity alone cannot guarantee.
An initial search can also carry editable smart defaults of at least 10,000 followers and a 0.5% engagement rate. Those settings sit well above the index’s 1,000-follower minimum. Unless a marketer notices and changes the filter chips, the results may create the misleading impression that suitable smaller creators are absent when they have simply been filtered out.
Super Search operates inside the brand’s collected library
Super Search solves a different retrieval problem. It analyzes individual images in an Archive workspace and finds UGC containing relevant products, objects, activities or visible attributes. The current Super Search specification, dated July 2, 2026, lists the tool on all five current plan categories and sets a maximum of 1,250 surfaced results in visual mode.
That displayed maximum is a retrieval cap, not a count of exact matches. Visual mode ranks items through meaning-based similarity, whereas caption and transcript modes return exact-match counts. Searching for a product shown in a frame and searching for its name in spoken or written text are consequently different operations and can produce different totals without either result being erroneous.
Carousel posts introduce another coverage gap. Super Search can return a carousel only when its first image or video matches the query; relevant material appearing solely on a later slide does not surface. An empty search result therefore cannot conclusively prove that the requested product or activity is absent from every collected carousel.
What the tools establish—and what they leave unresolved
Creator Search establishes that an account is a plausible visual or filtered match within Archive’s index. It does not prove sustained stylistic consistency, an appropriate audience, reliable performance or willingness to accept a particular campaign. Creator cards, contact fields and vetting lists organize the evaluation but do not replace it.
Super Search establishes that an asset appears relevant to a query within the limits of its ranking and supported content structure. Relevance does not establish permission to repost, edit or use that UGC in paid advertising. Origin, consent, licensing scope and campaign context remain separate questions even when the desired image is easy to locate.
Archive’s AI reduces search work rather than removing judgment. Creator Search produces a ranked and filterable pool of possible partners, while Super Search makes an existing UGC collection easier to navigate. The current documentation draws the boundary clearly: the software can identify plausible matches, but the brand remains responsible for qualification, rights review and decisions that depend on complete coverage.
Also read:
Subscribe to our newsletter
Get the latest Web3, AI, and crypto news delivered straight to your inbox.