What is Inillucent?
Inillucent is an embedded SQL database written in Rust for applications and agentic systems that need local data storage and search. It uses SQLite syntax and includes vector and keyword search for finding records by exact terms or meaning. The product also provides command-line, shell, and MCP tools, with builds for mobile apps and browser use. Its single database file supports access from multiple processes, while mobile and web versions can keep data on the device.
Inillucent Features:
SQLite SQL Support: Inillucent implements the SQLite dialect, including joins, recursive CTEs, triggers, foreign keys, indexes, and JSON functions, so existing SQLite queries and scripts can run as they are.
Vector and Keyword Search: The database supports vector similarity search, keyword search, and hybrid search, with an embed function that computes embeddings inside the database.
Confidence-Based Results: Search results include calibrated confidence scores, and results below a chosen threshold can be dropped so a query can return no matches.
Agent Tooling: A command-line interface, a shell with 63 of sqlite3's 65 dot commands, and an MCP server let users and agents work with databases.
Mobile and Web Runtime: Inillucent runs on iOS and Android and as WebAssembly in a browser, with tables and search indexes stored in the app or page.
Benchmark Performance: The product reports faster results than SQLite on 110 of 116 benchmark workloads and 500% faster overall.
Inillucent Benefits:
Reuse SQL Knowledge: Teams familiar with SQLite can build on established queries, schemas, and scripts instead of learning a different SQL dialect for their application.
Fewer False Leads: Confidence-based filtering can keep weak matches out of an agent's context, reducing the risk that it treats irrelevant passages as useful evidence.
Less Search Infrastructure: Keeping keyword and vector search in the database means a project can search locally without running a separate embedding service.
Local Data Access: Mobile and browser apps can search stored data without calling a server, which supports use when a network connection is unavailable.
Quicker Database Work: The reported benchmark results make Inillucent worth considering when query speed is a priority in an embedded database.
Use Cases:
Document RAG: Build retrieval for a corpus of Confluence documents, Slack messages, Jira tickets, or Figma designs using searches that match meaning as well as exact words.
Agent Database Access: Give an agent a way to query a database through MCP tools, such as when it needs to inspect records or retrieve information for a workflow.
Offline Mobile Search: Create an iOS or Android app that stores its tables, keyword index, and vectors on the device for searching without a server connection.
Browser-Based Catalogs: Put a searchable database in a web page, such as a film catalog, and keep its data in the browser between visits.
SQLite App Migration: Try Inillucent in an existing SQLite-based application when you want to retain SQLite queries and scripts while evaluating its search and performance.

