Analytics cookies only (Google Analytics, PostHog). No ad networks, no cross-site tracking. See our privacy policy.
An open-source search database for full-text, vector and hybrid search with real-time indexing, written in C++. You create tables and query them over SQL, either through a MySQL-protocol port or an HTTP JSON API, with faceting, filtering and columnar storage for datasets that do not fit in RAM. It is a database you call from your application, not a finished search page, and authentication for its SQL and HTTP ports is off unless you enable the `auth` setting.
Runs on a VPS from $6/mo (Kamatera 2GB, 2 GB RAM) · prices checked Nov 2026
Fits home hardware (26 MB measured idle, RAM only): Raspberry Pi 4 (4 GB) (arm64 untested), Raspberry Pi 5 (8 GB) (arm64 untested), Repurposed laptop (8 GB), Intel N100 mini PC (16 GB), AMD Ryzen mini PC (32 GB).
Pick Manticore Search when you want a search database you query with SQL and can size to data larger than RAM. It is a C++ engine for full-text, vector and hybrid search with real-time indexing, and the repository's LICENSE at the pinned tag is GPL-3.0 (the README badge says GPL v3 or later). You create tables and send queries either over a MySQL-protocol port (9306) or an HTTP JSON API (9308), and the README lists columnar storage for datasets that do not fit in memory. Our snippet adapts the Compose example in the project's Docker README, pinned to manticoresearch/manticore:29.9.0 (the 29.9.0 release is dated 2026-09-11), with the data directory on a named volume and both ports bound to loopback as upstream does. We did not find an upstream RAM minimum, so our 512 MB floor is an estimate. Meilisearch and Typesense are the nearest alternatives for application search with typo tolerance and per-tenant keys, and OpenSearch is the heavier log-analytics engine; Manticore sits between them as a SQL-first search database, and unlike SearXNG it is something your code calls, not a search page.
Paid link — we earn a commission if you shop through it.
Run these commands on your server:
# Manticore Search — adapted from the Docker-compose example in the manticoresoftware/docker README, pinned to 29.9.0
mkdir -p manticore && cd manticore
cat > docker-compose.yml <<'EOF'
services:
manticore:
image: manticoresearch/manticore:29.9.0
restart: unless-stopped
ports:
- 127.0.0.1:9306:9306
- 127.0.0.1:9308:9308
ulimits:
nproc: 65535
nofile:
soft: 65535
hard: 65535
memlock:
soft: -1
hard: -1
volumes:
- manticore-data:/var/lib/manticore
volumes:
manticore-data:
EOF
docker compose up -d
timeout 300 bash -c 'until curl -fsS -o /dev/null -X POST http://localhost:9308/sql -d "mode=raw&query=SHOW STATUS"; do sleep 3; done'
# → HTTP API on http://127.0.0.1:9308 and MySQL protocol on 127.0.0.1:9306 (authentication is off by default: keep both on loopback or a private network)Manticore's README and Docker documentation flag these before you rely on it:
Manticore Search used 26 MB of RAM at idle on GCP e2-standard-2 · Ubuntu 26.04 LTS · Docker 29.9.3 (first boot, measured Nov 2026). We list a minimum of 512 MB.
The Manticore Search install snippet is marked verified on Ubuntu 26.04 (Nov 2026).