How Instavision finds creators

Start broad, filter later

The biggest mistake in creator search is over-filtering at launch. Strict filters cut your output without cutting your effort — and you never see the borderline accounts that often turn out to be the best fits. Instavision flips this: at launch you set only what matters — the niche, the geography, and the audience-size band. Everything else is filtered after the results land, on a results page that works like a spreadsheet.

This is the single most important principle. You pay for the size of the scan, not the strictness of the query, so the right move is to cast a deliberately wide net and then curate.

Match on what creators actually write

Many of the best micro-creators never set a business category — but they mention their niche in their bio. So Instavision matches on bio language or category, not both. A forex mentor who writes “1-on-1 fx coaching” in their bio gets found even if their account type says “Personal blog.” Casting on the union of these signals is what surfaces the long tail other tools miss.

Run more than one discovery channel

No single method covers a niche. Keyword discovery reaches the long tail but misses peer-graph adjacencies; expanding from known leaders reaches peers but skews toward larger accounts. Instavision's smart playbook runs both and merges the results, which together cover the large majority of any niche's creator pool.

• •Cold-start broad — find a niche from zero using its vocabulary.
• •Peer expansion — map the network around creators you already rate.
• •Local creators — scope discovery to a city or region.

Narrow without rewriting your search

To drill from a broad theme (“wellness”) into a precise sub-niche (“sugar-free recipe creators”), you don't rebuild the query — you tighten a keyword whitelist. Keep the object-words that define the sub-niche (“recipe”, “sugar free”, “naturally sweetened”) and drop the role-words that pull in adjacent generalists (“coach”, “nutritionist”). The match happens during the scan, so off-topic accounts are filtered before they ever reach your list.

Iterate from broad to specific

Discovery is a funnel, not a single shot. A broad first pass surfaces candidates across adjacent themes; a tightened second pass narrows to your sub-niche; a peer-graph pass on the best finds adds deep-niche creators the first two passes missed. Each pass teaches you the vocabulary for the next.

Never pay for the same account twice

Iterating only works if runs don't keep returning the same names. Instavision remembers every account you've seen and, by default, skips them on the next run — so each search surfaces fresh creators instead of re-litigating your history. This was the number-one frustration we heard from people doing this by hand, and it's built in.

A relevance rubric you can apply in seconds

After the machine does its work, a quick human pass sorts the list. Instavision makes this fast by labelling every account with an AI category and a one-line reason, so you're confirming, not researching. The tiering we use:

The deliverable

Curate the keepers into a named list and export a clean PDF or CSV: clickable handles, full name, audience size, last-post recency, the AI category, and a one-line fit note per account — sorted and sectioned so it's ready to hand to a client or load into outreach.