August 2025 Summaries

Turso is an ambitious rewrite of SQLite in Rust, aiming for full compatibility while introducing features like Change Data Capture (CDC), concurrent writes, and encryption. The project, currently in alpha, encountered a peculiar issue where databases over 1GB appeared corrupted due to a missing special page that SQLite expects at the 1GB mark. This anomaly was identified and resolved by Nikita Sivukhin, a key team member whose approach and productivity have become legendary within the company. The discovery of the issue was facilitated by improving deterministic simulation testing, which initially failed to simulate databases growing beyond 1GB due to frequent fault injections. The SQLite source code's thorough documentation played a crucial role in identifying the problem, highlighting the challenges and intricacies of rewriting established software systems while maintaining compatibility.

Turso Database has introduced Model Context Protocol (MCP) support, enhancing the interaction capabilities between AI assistants and embedded databases like SQLite. MCP acts as a universal adapter, allowing AI models to connect to databases, APIs, and other systems seamlessly. By using a simple --mcp flag, Turso databases become accessible to AI assistants, enabling them to perform tasks such as querying data, modifying schemas, and executing database operations through natural language. This integration elevates SQLite databases to be more integral within the AI ecosystem, making them more versatile and user-friendly for tasks such as prototyping, data analysis, and database education. The MCP support allows AI assistants to assist in database design, suggest normalization strategies, and help refactor schemas, providing a more intuitive and efficient way to manage evolving project requirements. This innovation aims to streamline workflows by reducing the need for manual SQL coding and enabling more dynamic interaction with databases, fostering a more efficient development environment.

BÆRSkin Tactical Supply Co., a company known for durable outdoor gear, transitioned from using Kafka to SQLite for event ingestion, driven by the need to reduce complexity and address event loss issues. The company replaced Kafka's setup with a novel architecture where each Kafka partition is represented by a separate SQLite database, powered by Turso Cloud, allowing for infinite scalability and significant cost savings. This change resulted in a 60% reduction in infrastructure costs and eliminated event loss, leveraging the team's existing familiarity with relational databases. This approach challenges the conventional reliance on streaming platforms like Kafka, suggesting that SQLite-based solutions can offer efficient, scalable, and cost-effective alternatives for teams with a solid understanding of relational databases and moderate scale requirements.