All notable changes to this actor are documented here.
fetchAdDetails toggle: optionally retrieve every ad's full description text.
Leboncoin omits descriptions from search results entirely, so this is the only way
to get them. Billed at a separate, higher rate and only for ads whose description
was actually retrieved.raw field on every item — leboncoin's complete original response for that ad,
with nothing renamed or removed. Fields the simplified view leaves out are now
available: professional sellers' siren registration number, the full 17-field
location record (district, city label, region and department names), every image
size variant, has_phone, and each attribute's machine-readable value code
alongside its label.hasDetailData flag on every item, showing whether that ad's description was retrieved.- Automatic resume: interrupted runs (platform migration, timeout, manual abort)
continue from the last completed page instead of restarting — you are never charged
twice for ads a previous attempt already delivered.
- Tier-aware proxy selection: paying accounts automatically use a dedicated
network (measured ~2.6× faster with a lower block rate); free accounts use Apify's
residential proxy. No configuration needed.
- Tiered, usage-based pricing with a separate rate for ads including descriptions.
- Removed the proxy configuration input — proxy selection is now automatic based on
your account plan, and there is nothing to set up.
- Existing output fields are unchanged.
raw and hasDetailData are additions, so
existing integrations and dataset views keep working exactly as before.
- Duplicate ads are no longer charged twice. Leboncoin repeats ads across pages
and overlapping searches; these are now detected and skipped (a 1,000-ad test run
caught 7 duplicates that would previously have been billed).
- Ads returned by a search are no longer missing the description field silently — it
is either filled (with
fetchAdDetails) or explicitly flagged via hasDetailData.
- Invalid or unparseable start URLs now return a clear explanation in the dataset and
let the run finish successfully, instead of failing the whole run.
- Runs no longer waste time on a stale site build reference; the correct one is
resolved automatically at startup.
