> ## Documentation Index > Fetch the complete documentation index at: https://docs.lobstr.io/llms.txt > Use this file to discover all available pages before exploring further. # November 2, 2026 > Slots: only runs in progress use them; squids capped at 100 per slot **November 2, 2026** Google Maps Leads Scraper: geo\_match keeps results within 10 km, everywhere ## Improved **Google Maps Leads Scraper: `geo_match` keeps results within 10 km of the search point, in every country** `geo_match` now has a single rule: a result is kept when it lies within 10 km of the search point, which is the coordinates of the task's city, or the `@lat,lng` center of a Google Maps URL. Postal codes are no longer checked, and Google Maps URLs go from 50 km to 10 km. The postal code rule dropped real places: a postal code often spans several towns, business-only codes (French `CEDEX`) never match a place, places that hide their address have no postal code, and a search on one city lost the business 2 km away in the next town. On recent runs, 96.8% of the places kept were within 10 km of a search point. To cover a large area, add one task per city. See [Update Settings](https://docs.lobstr.io/examples/google-maps-leads-scraper/update-settings.md). **Google Maps locations: one city per point on the map** [Get Geolocation](https://docs.lobstr.io/examples/google-maps-leads-scraper/get-geolocation.md) now returns one city per coordinates. Several postal codes of a town often share the town's coordinates (business-only `CEDEX` codes in France, several codes per town in Portugal), so a task per entry ran the same search again: 28% of the French entries and 43% worldwide. The list of a region gets shorter (Provence-Alpes-Côte d'Azur: 2,190 to 997 cities) for the same results. City names you already use in tasks keep working. **November 2, 2026** Slots: only runs in progress use them; squids capped at 100 per slot ## Improved **Slots: only runs in progress use them, so you can create as many squids as you need** A squid used its `concurrency` in slots for as long as it stayed active, running or not: on the Free plan (1 slot), creating a second squid answered `SlotsLimitExceeded` until the first was deleted or deactivated. Slots now count the runs in progress only. Creating or activating a squid uses no slot, and a plan holds up to 100 squids per slot (Free 100, Starter 200, Pro 1,000, Team 5,000); past that, [Create Squid](https://docs.lobstr.io/docs/create-squid.md) answers `SquidsLimitExceeded`. A launch is refused with `SlotsLimitExceeded` when your other runs in progress already use your plan's slots, and nothing is queued: launch again when a run finishes, or stop one. The same applies to resuming, resurrecting or retrying a run, and to raising a running squid's `concurrency`. A scheduled run due while your slots are busy ends with the `slots_limit_exceeded` done reason. [Get Balance](https://docs.lobstr.io/docs/get-balance.md): `used_slots` now counts the slots of your runs in progress, and two fields are new, `squids` and `max_squids`. Your existing squids and schedules behave as before: you never had more active squids than slots, so your runs never used more than your plan's slots. **November 2, 2026** Instagram Comments Scraper: latest comments only, no sort order, no replies ## Breaking **Instagram Comments Scraper: latest comments only, no sort order, no replies** Instagram no longer gives visitors without an account a sort order, replies or reply counts, and it serves only a post's most recent comments: all of them on a small post, the last few hundred on a busy one. The scraper works without an account, so: * `sort_order` and `include_child_comments` are removed from the [Update Squid](https://docs.lobstr.io/docs/update-squid.md) parameters. Sent anyway, they are ignored. * `replies_count` is `null`, `is_child_comment` is always `false` and `parent_comment_id` is always empty. ## Fixed **Instagram Comments Scraper returns comments again** Since about October 7, every run ended with `invalid_url`: Instagram had closed the address the scraper used to read comments. It now reads them the way instagram.com's own post page does. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.