valeriansaliou/sonic

Health v2 · maintenance only

Adoption not part of the score

21321 stars · 618 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Sonic is a fast, lightweight, schema-less search backend written in Rust that ingests search texts and identifier tuples, returning matching IDs for queries in microseconds. It serves as a minimal alternative to Elasticsearch, running on ~30MB of RAM with auto-completion and natural-language query normalization.

Use cases

• add full-text search to my app without running elasticsearch
• lightweight search backend that runs on a small server
• autocomplete and search suggestions for my website
• index millions of records with low memory usage
• search backend that returns document ids instead of storing documents
• schema-less search server for a small project

When to choose

• you need fast search with minimal RAM and CPU footprint
• you only need an identifier index backed by an external database
• you want simple search and autocomplete without Elasticsearch's complexity
• you're running on a small, cheap server

When to avoid

• you need full-featured features like aggregations, faceting, or complex queries
• you need document storage and retrieval from the search engine itself
• you need distributed clustering and horizontal scaling out of the box

Facets

service · maturity stable

search-engine http-server database backend developer-tools databases rust self-hosted cross-platform schema-less elasticsearch-alternative full-text-search autocomplete lightweight identifier-index search linux macos docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="valeriansaliou/sonic")