# Changelog of LinkedIn Person Search | X-Ray | No Login | Bulk (`aeonkosmos/linkedin-person-search`) Actor

- **URL**: https://apify.com/aeonkosmos/linkedin-person-search/changelog.md
- **Full Actor documentation**: https://apify.com/aeonkosmos/linkedin-person-search.md

## Changelog

### 2026-09-18

#### Pricing model — per search query AND per result

- Added a **search query** event, charged once per search executed, whether or not it
  returns results. Previously an empty or low-yield search billed the full upstream cost
  and earned nothing.
- **Result** event rebalanced alongside it.
- Removed the claim *"No results = no charge"* from the Store description. It was
  inaccurate: a search that finds nothing still costs a query.
- See the Pricing tab for current rates.

#### Permissions — moved to Limited

- Runs under **limited permissions**. The actor only needs its own storages and public
  profile information. This removes the manual approval prompt for new users (including
  MCP, CLI and API runs) and improves the Actor Quality score.

#### Performance and cost

- Default memory reduced from 4096 MB to **512 MB**. The actor performs HTTP requests and
  JSON parsing only; compute is memory x time, so this is an ~8x reduction in run cost.

#### Reliability

- Search loops now abort immediately when a run's spending cap is reached, instead of
  continuing to issue searches that cannot be billed.
- Input schema trimmed to the modes that are actually implemented, so invalid input is
  rejected at run creation rather than silently doing nothing.

#### Listing

- Added SEO title and description.
- Added a real example input (previously a placeholder).