Danube Messaging

Open-source pub/sub and streaming platform built in Rust for reliable messaging without the operational overhead. One self-contained binary with embedded consensus, runs standalone for dev, scale to a cluster, or deploy at the edge.

Get Started View on GitHub

Try Danube

Download the latest binary from the releases page or pull the container image.

Start a broker and test with the CLI:

# Start the broker
danube-broker --mode standalone --data-dir ./danube-data

# Produce messages
danube-cli produce -s http://127.0.0.1:6650 -t /default/demo -c 10 -m "Hello, Danube!"

# Consume messages
danube-cli consume -s http://127.0.0.1:6650 -t /default/demo -m my_sub

Or run a single broker in a container:

docker run -p 6650:6650 -p 50051:50051 \
  ghcr.io/danube-messaging/danube-broker:latest \
  --mode standalone --data-dir /data \
  --broker-addr 0.0.0.0:6650 --admin-addr 0.0.0.0:50051

Run the following command on each of your 3 nodes to form a Raft consensus cluster:

danube-broker --config-file danube_broker.yml \
  --broker-addr 0.0.0.0:6650 --raft-addr 0.0.0.0:7650 \
  --data-dir ./data/raft \
  --seed-nodes "node1:7650,node2:7650,node3:7650"

Or use Docker Compose to start a 3-node cluster with Prometheus monitoring:

git clone https://github.com/danube-messaging/danube.git
cd danube/docker/quickstart
docker-compose up -d

Check cluster health:

danube-admin brokers list
danube-admin brokers balance

MQTT gateway for IoT devices at the edge. Validates payloads against schemas, buffers into a local WAL that survives outages, and replicates to your central cluster over gRPC.

danube-broker --mode edge --data-dir ./edge-data --edge-config edge.yaml

Why Danube?

One Binary

Run as Standalone for development, Cluster for production, or Edge for IoT ingestion, all from the same danube-broker binary with no external dependencies. Learn more →

Pub/Sub & Streaming

Publish to topics, consume via named subscriptions. Use best-effort dispatch for fire-and-forget real-time workloads, or persistent dispatch with WAL-backed at-least-once delivery, acknowledgments, and dead-letter queues. Learn more →

Resilient Clusters

Embedded Raft consensus with automatic leader election, failover and topic reconciliation on restart. Intelligent load manager monitors broker health and rebalances topics to prevent hotspots. Scale horizontally with zero downtime. Learn more →

Cloud-Native Storage

Local WAL for fast writes with background export to S3, GCS, or Azure Blob. Tiered reads serve from cache when possible, falling back to durable segments. Topics move between brokers without data loss. Learn more →

Schema Registry

Built-in schema management with JSON Schema, Avro, and Protobuf. Version control with backward, forward, and full compatibility enforcement, no separate registry service to deploy. Learn more →

AI-Native Admin

Manage your cluster with natural language via the Model Context Protocol. 58 tools and 4 guided workflows covering health checks, schema evolution, security setup, and rebalancing through Claude Code, Antigravity, Cursor, or any MCP-compatible assistant. Learn more →

Explore

Client Libraries Rust, Go, Java, and Python SDKs with producer, consumer, and schema registry examples.

Security TLS/mTLS encryption, JWT authentication, and fine-grained RBAC with default-deny semantics.

Edge & IoT MQTT gateway with schema validation at the edge, WAL-buffered replication to the central cluster.

Danube Iceberg Continuously export streaming data to Apache Iceberg tables, queryable by Spark, Trino, DuckDB, and any Iceberg-compatible engine.

Client Libraries

Rust · crates.io · examples | Go · pkg.go.dev · examples | Java · Maven Central · examples | Python · PyPI · examples