Search platforms (non-memory)
8 systems in the search platforms (non-memory) category of the AI Agent Infrastructure Landscape, grouped by maturity tier.
Tier 1 — battle-tested (3)
- Algolia Hosted search API
Search-as-a-service API for instant search experiences — NeuralSearch (vector) + lexical. Adopted by 12000+ companies including LVMH, Stripe.
- OpenSearch Open Elasticsearch fork
AWS-led fork of Elasticsearch 7 — Apache-licensed; k-NN plugin, OpenSearch ML, neural-search. Substrate beneath AWS OpenSearch Service.
- Supabase Vector Postgres + pgvector cloud
Supabase's managed pgvector — combined Postgres + vector store; popular default for OSS LLM apps. Supabase $80M Series C 2024.
Tier 2 — production-ready (3)
- Quickwit Log + event search engine
Open-source distributed search engine for logs/traces — cloud-native (object storage as primary tier). Acquired by Datadog 2024.
- Tantivy Rust full-text search library
Rust port of Lucene — embedded full-text search engine. Underlies Quickwit, Lnx, Bayard. Used in Meilisearch and many in-process search engines.
- Typesense Open-source typo-tolerant search
Open-source search engine focused on typo-tolerance and ease of use — Typesense Cloud commercial; rising Algolia alternative.
Tier 3 — emerging (2)
- Sajari (now Search.io) Site search platform
Site-search platform (formerly Sajari) with the Neuralsearch vector engine. Acquired by Algolia in September 2022 and folded into Algolia's hybrid search; search.io now redirects to algolia.com.
- Vald Cloud-native distributed vector search
Yahoo Japan's open-source distributed vector search engine — built on NGT; Kubernetes-native; horizontal scaling.