PlanetScale contributors Eric Ridge and Patrick Reynolds launched TIN (Text INdex) on September 16, 2026 — a GA full-text search extension for Postgres and Neki. Queries use the ==> operator with TINQL supporting Boolean, phrase, span, fuzzy, wildcard, regex, BM25 ranking, and MVCC-correct updates.
On a 150M-document corpus, PlanetScale benchmarks report index build 2–16× faster than alternatives and up to 541× higher query throughput vs GIN, using ctid-based postings and vectorized bitmap operations.
Lead, an AGPL companion extension, accepts TIN-compatible SQL for local dev/CI but deliberately scans tables and is unsuitable for production. TIN is PlanetScale-managed and not self-hostable outside their cloud products as of September 2026.