v4.0.0

4.0.0 (2026-09-22)

⚠ BREAKING CHANGES

• search.companies(company=...) no longer type-checks a dict literal containing linkedin_url. The endpoint never honoured it, so drop the key. On search.people and company.tam_by_people it keeps working, now typed as PeopleCompanyFilter (exported from blitz_api and blitz_api.types).
• Company.slogan, Company.revenue and Company.employee_growth are removed, as is the EmployeeGrowth model and its blitz_api.types export -- the API never returned any of them. Company.specialties is now list[str] defaulting to [] rather than list[str] | None; code branching on specialties is None should check for an empty list instead.
• InsufficientCreditsError is removed. Catch InsufficientRecordsError instead.
• Education.field_of_study is removed — the API folded it into degree ("Bachelor of Science, Industrial Engineering").

Features

• add the Industry.UNKNOWN bucket and document the range-filter 422 (5f2d944)
• remove the deprecated InsufficientCreditsError alias (0c9be23)
• split PeopleCompanyFilter out of the shared CompanyFilter (fc6983e), closes #25
• sync SDK to the 2026-09-11 / 2026-09-15 API releases (81460ed)

Bug Fixes

• remove phantom Company fields and the redundant _drop_none layer (52b5ec5)

Documentation

• correct the profile_min_connections default to 0, not 200 (6ee2893), closes #26
• correct the README headline example and route to enrichment.person (7051afb), closes #30
• correct the stale rate-limit and spec-version facts (10615d5)
• correct the stale response-schema rule and link the JS mirror PR (1f1381f)
• guard the optional linkedin_url in the README enrichment example (cdcfb71)
• name the legacy 50 req/s rate limit instead of "more on legacy ones" (f4974a1), closes #29
• the /changelog/ trailing slash is not load-bearing (891d5b7), closes #28