Honker, a new SQLite extension, lets developers add Postgres‑style NOTIFY/LISTEN semantics to a single database file. The Rust‑based crate ships with bindings for Python, Node, Bun, Ruby, Go, Elixir, and C++. By converting polling into WAL‑based event notifications, it delivers single‑digit millisecond cross‑process reaction time in production, avoiding separate brokers and teams.

SQLite dominates embedded and mobile projects, yet most need task queues or pub/sub. Typical solutions add Redis and Celery, doubling the datastore and backup burden. Honker keeps the queue inside the same file, so an INSERT into orders and queue.enqueue() share a transaction—commit or rollback together—eliminating dual‑write issues. This approach simplifies backups, reduces operational overhead, and ensures atomicity across business data and work items, making it an attractive alternative for developers who prefer a single, lightweight database stack today.

Honker offers durable work queues with retries, priorities, delayed jobs, and a dead‑letter table. It supports crontab‑style periodic tasks, named locks, rate limiting, and task result storage via enqueue IDs. Streams provide per‑consumer offsets and configurable flush intervals, while a notify API delivers transient messages without replay. All exposed through a single SQLite loadable extension.

In practice, developers can open honker with a single call, enqueue jobs atomically with business writes, and claim them in workers that react to WAL commits. The extension’s API mirrors familiar patterns from Celery or Huey, lowering the learning curve. By consolidating data and queueing, honker cuts infrastructure complexity and improves reliability for production teams.