Inillucent

Inillucent, from Black Rainbow Labs, is positioned as an embedded SQL database for agentic systems. It is written in Rust and implements a complete SQLite dialect, so existing SQL knowledge and SQLite-style workflows carry over. A single database file supports multi-process access, and the project emphasizes a battle-tested log and recovery mechanism.

The feature set centers on speed and search. The site claims faster performance than SQLite on 110 of 116 benchmark workloads and roughly 500% faster overall, with notable gains on key reads and from Python, Node and the shell. Vector embeddings for semantic search and keyword search are built in rather than bolted on, and the site reports large advantages over Postgres and Postgres plus pgvector on those search workloads.

Tooling includes a full-featured command line interface, an interactive shell and an MCP server, which suits AI agents that need to query data directly. A Postgres dialect option is listed as upcoming, so current compatibility is SQLite-oriented.

It fits developers building agentic or AI applications who want an embedded database with semantic search and low-latency local access, especially those already comfortable with SQLite syntax. Teams deeply invested in Postgres-specific features should wait for or verify the Postgres dialect. A sensible approach is to run the published benchmarks against your own workload and test the MCP integration before adopting it in production.