Unix pipes for blockchain indexing. No complex infrastructure required.
Written in Go. Decouples fetching from processing. Backfill 5 years of history in hours, not months, by parallelizing fetch workers.
No heavy databases required. nebu respects stdin and stdout. Pipe directly into DuckDB for instant SQL analytics, jq for filtering or any tool that you want.
Bridge the gap between CLI and Cloud. Use nats-sink to turn your local pipeline into a distributed, real-time firehose for your API.
| Pipe anything. | Replace any component. | Build what you need. |
Get nebu installed on your system in seconds
Requires Go 1.23 or later. After installation, the nebu command will be available in your $GOPATH/bin.
Install processors on-demand using nebu install <processor>
Common commands to get you up and running with nebu
Separate fetch from processing
GCS/S3 for data lakehouse
Use authenticated endpoints
Extract → filter → dedupe → store
List and install components
Stream events from Stellar ledgers
USDC transfers only
Like tail -f
Fan-out with tee
Aggregate with DuckDB