set up olywork — https://olywork.com/llms.txt
Using olywork, screen these 20 Instagram creators for partnership fit: [list of handles]. For each one pull the profile (followers, category, is_private, business_email) and the last 12 posts. Show me the price before each step. If a business email is public, verify it. Give me a CSV with handle, followers, private, posts_sampled, avg_engagement_pct and a keep or skip column: keep if over 10k followers, public, posting, and engagement above 1%.
The agent runs one profile call and one posts call per handle. One batch beats twenty prompts.
Both TikHub routes bill per success at the rate in the table; the verify step is the one that costs, and only fires when an email exists.
Follower floor, public account, recent posts, engagement floor. The agent filters; you set the rules.
Structured output you can sort before spending time on outreach.
Prices are the provider’s own rate, read from the catalog when this page loads, with $0.000 added by olywork.com. Success rates are olywork.com’s own served calls over the last 30 days.
| # | Step | What the agent asks | Provider used | Price | Success rate |
|---|---|---|---|---|---|
| 1 | Pull the profileThe cheapest profile route in the catalog. Returns the numeric user id the posts step needs. | followers, category, private flag and public business email, by handle | $0.001 per found | not yet measured | |
| 2 | Sample recent postsEngagement is likes plus comments over followers, averaged across the sample. Private accounts skip this step. | the last 12 posts with likes and comments, by numeric user id | $0.001 per found | not yet measured | |
| 3 | Verify the business emailFires only for handles whose profile exposes an email, so the worst case in the total above rarely happens. | is the public business email deliverable, before outreach | $0.0123 per call | not yet measured |
Run on 2026-09-14, 20 creators in.
Thirty-six successful calls at the two TikHub rates, billed per success. Fourteen of the twenty made the keep list. None of these creators exposes a business email in the profile, which is common: the email step is there for the lists where they do, and it is the only step with a real price. One profile call did not answer and was not billed; the retry policy is one line in the prompt. Handles are removed from the published CSV because they identify real people; your own run returns them.
olywork.com holds the provider keys. Neither you nor the agent sees them.
The provider's own rate, $0.000 markup, from a prepaid balance.
Charged per call. $1.00 free per new team, no card to start.
Already pay Hunter? Register it and those calls are never metered.
Another provider is a different word in the prompt, not a new integration.
No SDK, no OAuth dance per vendor, no seats.
The profile call returns followers, bio and the private flag; the posts call returns nothing. Decide keep or skip on the profile alone.
Most creators do not expose one. The profile returns what is public; the verify step is skipped, not billed.
A renamed handle returns no user. The miss is free on this route; fix the handle and rerun that row.
Spread a few hundred handles over minutes, or the profile route starts answering slowly.
Do not point a logged-in personal or brand Instagram session at bulk pulls. Use a public-data route that does not need your cookies. If a provider asks for a session, use a throwaway account you can lose.
Bought followers and engagement pods still show up as healthy looking profiles. Sample recent posts and compute engagement before you put someone on an outreach list.
Opening profiles one by one is the common starting point and it breaks past a few dozen handles. Batch the profile and posts calls, then decide keep or skip in the CSV.
The receipt above is one real run. Profiles and posts are the cheap steps; only the email verify step, when it fires, costs more.
The profile call still returns followers, category and the private flag. Posts and engagement are unavailable.
Yes. TikHub has TikTok profile and post routes at the same price; change the handles and the platform in the prompt.
Likes plus comments on the sampled posts, divided by followers, averaged over the sample. Above 3% is generally good; this list's median was 5.3%.
This workflow screens public profile and post fields through a metered catalog route. It is not a follow, unfollow, or DM bot. Do not paste your own Instagram session into a scraper.
Treat engagement as a filter, not a guarantee. Look at comment quality and sudden follower spikes before you pay for a placement.