Find long-running Facebook ads in your niche.

Confirm this controlled long-running Meta ad scope: automotive in US. Use active ads, minimum launch age 30 completed UTC days and historical floor 2020-01-01. Exclude the open UTC date and let `L` be the latest completed UTC day. Calculate exact inclusive buckets as 30-60=`L-60..L-30`, 61-90=`L-90..L-61`, 91-365=`L-365..L-91`, 366+=`2020-01-01..L-366`. Do not call tools. Return only the validated scope and dates. Stop after Step 1.

Cost guard: before each paid Spytrend call, run the free `estimate_cost` with the same tool and arguments, estimate the cost and stay within the user's stated budget; if the user has not stated a budget, tell them the estimate before the first paid call. If pricing consent is required, show the returned terms and stop until the customer explicitly agrees; never accept on the customer's behalf. If a required call does not fit the stated budget, make no paid call and return `Budget limit prevents required research`; omit an optional call that does not fit and disclose reduced coverage. When opening ads with `get_ad`, skip multilang ads (`is_multilang`) and never call `get_media` after `get_ad`. Retry once only when `retryable=true`, after `retry_after_seconds`. Read `params.ignored` after every search; an ignored required filter never counts as applied. If a search reports `pagination.total_status=restricted` (a free-plan archive window; `meta.plan_window_cutoff` names the boundary), tell the user the result covers only ads that started on or before that date, and never report a zero from it as "no such ads exist". Cost notices are not part of the final customer report.
Run four search_ads routes sequentially for the retained buckets. Keep source Meta, exact category/subcategory, selected country, active status, dedupe=true, date ascending and limit=3 fixed. Require explicit date_from/date_to and confirm parameter application. A completed zero is valid. Retry exactly once only after explicit retryable=true and retry_after_seconds, with a fresh estimate before the identical retry. A failed bucket is unavailable and makes the run Partial; continue the others. Return only bucket outcomes and candidate counts. Stop after Step 2.

Cost guard: before each paid Spytrend call, run the free `estimate_cost` with the same tool and arguments, estimate the cost and stay within the user's stated budget; if the user has not stated a budget, tell them the estimate before the first paid call. If pricing consent is required, show the returned terms and stop until the customer explicitly agrees; never accept on the customer's behalf. If a required call does not fit the stated budget, make no paid call and return `Budget limit prevents required research`; omit an optional call that does not fit and disclose reduced coverage. When opening ads with `get_ad`, skip multilang ads (`is_multilang`) and never call `get_media` after `get_ad`. Retry once only when `retryable=true`, after `retry_after_seconds`. Read `params.ignored` after every search; an ignored required filter never counts as applied. If a search reports `pagination.total_status=restricted` (a free-plan archive window; `meta.plan_window_cutoff` names the boundary), tell the user the result covers only ads that started on or before that date, and never report a zero from it as "no such ads exist". Cost notices are not part of the final customer report.
Deduplicate exact ad IDs and skip multilang rows (`is_multilang`). Open at most six candidates in total, oldest first, at least one from every non-empty bucket and no more than two from one bucket, preserving advertiser diversity; estimate each opening and stop opening once five ads are verified. Verify detailed active status, selected-country delivery, explicit Facebook start date inside its bucket, requested format when controlled and direct offer fit. Taxonomy alone is insufficient. Do not run replacement discovery. Return only verified count, source buckets and Complete/Partial. Stop after Step 3.

Using retained evidence only, write a table-free customer report of no more than 450 words with 3–5 verified ads (fewer only when fewer passed verification, with the reason). Show the category by its human name, never a slug. For each ad show exact start date, observed launch age, status, offer, advertiser, format/CTA, landing page, one reason to inspect, all four bucket outcomes and coverage limits. Hide service name, calls, raw IDs, tokens, timing, retries and run summaries. Do not rerun research. End exactly: `Launch age and current active status do not prove continuous delivery, spend, conversions, revenue, ROAS or profitability.` Stop after Step 4.