Verified Content Keys
Content-defined chunking and cryptographic hashing ensure every blob is stored and addressed by its bytes.
Graviton
Content-Addressable Storage Runtime
Built on ZIO • Streaming ingest • Backend agnostic
Skip to content 
Operations Snapshot
Track ingest throughput, deduplication ratios, replica health, and runtime events directly in the docs. The live HUD mirrors the data surfaced by the runtime's Prometheus exporters so you can see what operators monitor day to day.
HUD
Telemetry Feed
Real-time runtime signalsQuick Start
bash
# Build all modules
./sbt compile
# Build the Scala.js dashboard for the /demo page
./sbt buildFrontend
# Run the full test suite
TESTCONTAINERS=0 ./sbt scalafmtAll test
# Launch this documentation site
cd docs && npm install && npm run docs:devWhy Graviton?
Graviton is a content-addressable storage runtime that coordinates chunking, hashing, replication, and retrieval for large binary payloads. Each concern lives in an isolated module so hashing algorithms, network protocols, and storage backends can evolve independently.
Visualize the pipelines
Architecture, manifests, and operations pages include interactive Mermaid diagrams rendered client-side in VitePress. Follow ingest, replication, and backend selection without leaving the browser.
Key Features
- Composable Transducer pipelines — build ingest, verify, and retrieval paths from typed stages with
>>>and&&& - Streaming ingest and retrieval with ZIO Streams and zero-copy pipelines
- Content-defined chunking via FastCDC and multi-hash verification
- Pluggable storage backends for S3, PostgreSQL, RocksDB, and future targets
- Strongly-typed schemas shared across HTTP, gRPC, and Scala.js clients
- Integrated observability with Prometheus metrics and structured logging
- Replica coordination through policies that balance durability and latency
- Interactive Pipeline Explorer — compose and visualize transducer stages in the browser
Pipeline Explorer
Compose transducer stages interactively — toggle stages on and off, see the composition expression update in real time, and watch animated data flow through the pipeline.
Transducer Pipeline Explorer
Compose stages with >>> and &&& — watch data flow in real time
Available Stages
CountBytes >>> HashBytes >>> Rechunk#Count Bytes
Chunk[Byte] => Chunk[Byte]Total0 B
idle
>>>
HHash Bytes
Chunk[Byte] => Chunk[Byte]Digest---
Hashed0 B
idle
>>>
RRechunk
Chunk[Byte] => Chunk[Byte]Blocks0
Fill0%
idle
yield
Summary
totalBytes0 BLong
digestHex---String
hashBytes0 BLong
blockCount0Long
rechunkFill0%Int
Data Flow
3x
Raw Bytes Block Hash Manifest
Try a scenario
Open full Pipeline Explorer for detailed explanations and scenarios.
Next Steps
New to Graviton?
Start with the Getting Started Guide for a hands-on introduction!
CAS Playground
Type text and watch it get chunked, hashed, and deduplicated in real time
Pipeline Explorer
Compose transducer stages interactively and watch data flow in real time
Architecture
Understand the module-by-module breakdown and system design
Upload Flow
Follow a binary blob through the entire ingest pipeline
Transducer Algebra
Typed, composable pipeline stages with Record summaries
API Reference
Explore gRPC and HTTP endpoints with examples
Contributing
Join the community and help build the future of storage