Manticore Search - Full Text and Vector Database

This is a repackaged open source software product wherein additional charges apply for cloudimg support services.

## A Search Database You Run Yourself

Manticore Search is an open-source search database for full text, vector and hybrid search, with real-time indexing and a SQL interface. It is the modern successor to Sphinx and is built for workloads a general-purpose database handles badly: relevance-ranked text retrieval, faceted product catalogues, log and document search, and semantic retrieval over embeddings.

This cloudimg image delivers the engine preinstalled, hardened and answering queries within minutes of launch on any standard EC2 instance, with no cluster to provision and no per-document indexing fee.

## About cloudimg

cloudimg publishes production-hardened machine images on AWS Marketplace, each built with security-first defaults so teams can adopt open-source infrastructure without the configuration risk. Our 24/7 support team covers deployment, schema design, performance tuning and ongoing operations.

## Closed by Default, Not Open by Default

A stock Manticore installation leaves its SQL and HTTP interfaces unauthenticated because upstream expects them on a trusted network. Shipping that arrangement to the internet would hand anyone who finds the instance full read and write access to your index. This image does the opposite:

• Every Manticore listener is bound to loopback only and none is opened in the security group
• The HTTP interface is published solely through an nginx reverse proxy behind an HTTP Basic Auth gate
• The gate password is 24 characters, generated uniquely on each instance at first boot, stored bcrypt-hashed, and never baked into the image
• An unauthenticated health endpoint is provided for load balancer probes and exposes nothing about your data

No two instances launched from this image share a credential, and no default password ships.

## Search Types in One Engine

• Full text search with BM25 relevance ranking, phrase and proximity operators, stemming and stop words
• Vector search over float vector attributes with approximate nearest-neighbour indexing for semantic retrieval
• Hybrid search that fuses a keyword result list with a vector result list, so exact identifiers and loose paraphrases both land
• Faceted aggregation that counts attribute buckets in the same round trip as the query
• Percolate queries that store queries instead of documents, for alerting and saved-search notifications

## SQL and HTTP - No New Query Language

Manticore speaks the MySQL wire protocol, so existing clients, drivers and BI tools connect directly and issue ordinary SELECT, INSERT, UPDATE and JOIN statements. The same SQL is available over HTTP, which is how this image publishes the engine through its authenticated proxy, alongside a JSON API for search, bulk insert and index management.

## Ready to Use on AWS

The search index lives on its own dedicated EBS volume, mounted by filesystem identifier, so a growing corpus never competes with the operating system for space and can be resized independently. A demo table is indexed in the image so you can run a real relevance-ranked query on first boot before loading any data of your own. Everything runs under systemd on Linux hosts, fully patched at build time.

Open port 22 for SSH and port 80 for the search API in your security group and you are operational. An m5.large is a comfortable default; search is memory-sensitive, so size up as your working set grows.

## Key Use Cases

• Site and catalogue search: Relevance-ranked product, document or article search with facets, running on infrastructure you control.
• Log and event search: Real-time ingestion with immediate searchability and no offline rebuild window.
• Semantic and hybrid retrieval: Vector similarity beside keyword matching in one engine, without bolting a separate vector store onto your stack.

## Evaluate Before You Commit

A 7-day free trial is enabled on this listing. Launch, retrieve the generated password over SSH, query the demo table through the authenticated endpoint, then load a slice of your own data and measure it. When you are ready for production, scale the instance and grow the index volume.

## 24/7 cloudimg Support

Our team provides 24/7 technical support by email and live chat covering deployment, putting TLS in front of the search endpoint, schema and index design, relevance tuning, real-time versus plain table choices, vector and hybrid search setup, backup with the bundled tooling, and sizing guidance.

Manticore Search is distributed under the GNU General Public License version 3 or later. All product and company names are trademarks or registered trademarks of their respective holders. Use of them does not imply any affiliation with or endorsement by them.