Fun things I've built for myself and others.

Open Source

hadb Cheap HA embedded databases via cloud storage. No custom failover logic
haqlite HA SQLite with leader election, WAL replication, and write forwarding via S3
hakuzu HA Kuzu/LadybugDB with leader election, journal replication, and write forwarding via S3
graphstream Journal replication engine for graph databases. Ships logical WAL to S3
haduck DuckDB extension for block-level replication to S3 via custom FileSystem
duckblock Block-level replication primitives for DuckDB. Ships dirty 256KB blocks to S3
hadb-changeset Unified replication formats for the hadb ecosystem. Binary wire format for shipping pages over S3
turbolite SQLite VFS for S3-tiered storage with compression and encryption
sqlite-tantivy SQLite extension for Tantivy-powered full-text search with FTS5-compatible API
forkstack Instant, isolated dev environments using zero-copy database and storage forks
tenement Process hypervisor for single-server deployments. Pack 100+ services on a $5 VPS
honker SQLite extension adding Postgres-style NOTIFY/LISTEN with durable pub/sub and task queues, no broker
redb-turbo Fork of redb with AES-256-GCM encryption and zstd compression
sqlite-loadable-rs Framework for building loadable SQLite extensions in Rust (fork with bugfixes)
dash-security Building blocks for secure modern Dash apps with Dash Pages
brain-plasma Shared inter-thread memory for Python webapps
dash-access Access control for Dash apps
dash-auth-flow Auth for Dash apps
duffel Tiny Python dataframes without Pandas/Numpy (e.g. AWS Lambda)
dash-slides Interactive presentation slides in Python