DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///
DARC.DEV
Tool radar
p

Retrieval · Documentation review

pgvector

Vector similarity search alongside relational data in PostgreSQL.

pgvector adds vector types and similarity operators to PostgreSQL. Exact nearest-neighbor search is the default; HNSW and IVFFlat indexes trade recall for speed. It is attractive when relational metadata, transactions and existing database operations are central to the application.

Official documentationReviewed 23 Sept 2026
Source-checked Sources & corrections

A good fit for

  • → PostgreSQL-based applications
  • → Joining vectors with ownership and metadata
  • → Small-to-medium retrieval systems before specialization

Know the limits

  • → Approximate indexes change result behavior.
  • → Selective filters can interact with ANN recall.
  • → Embedding dimensions, distance metric and index operator class must agree.

Your first useful workflow

01

Establish exact recall

Keep a small reference query set and an exact-search baseline.

02

Add one index

Measure latency and relevance before tuning several knobs at once.

03

Test filters

Evaluate the actual permission and category predicates used in production.

The practical take

Start with exact search on a small corpus before paying the complexity cost of ANN tuning.

Sources & evidence

Capabilities above are summarized from these official documents. Workflow suggestions are editorial guidance, not a claim of measured product superiority.