Camunda 8 Run
«Camunda Run is a packaged, single-application distribution of Camunda 8 designed specifically for local development and testing— not for production.»
- Download Desktop Modeller
- Download Camunda 8 Run:
This version for example: https://downloads.camunda.cloud/release/camunda/c8run/8.9/
Extract, and then:
Run the helper script: ./start.sh
Start Camunda Run: ./c8run start
Stop Camunda Run: ./shutdown.sh
when running ./start.sh:
francd@k-legion:~/c8run-8.9.0$ ./start.sh
9:25PM INF Resolved secondary storage type from configuration config=/home/francd/c8run-8.9.0/configuration/application.yaml secondaryStorage.type=rdbms
System Version Information
--------------------------
Camunda Details:
Version: 8.9.0
--------------------------
Java Details:
Version: 25.0.2
JAVA_HOME: /home/francd/.sdkman/candidates/java/25.0.2-tem
JAVA_OPTS:
--------------------------
Logging Details:
Connectors: ./log/connectors.log
Camunda: ./log/camunda.log
--------------------------
Press Ctrl+C to initiate graceful shutdown.
--------------------------
9:25PM DBG Failed to read PID from file. This is expected for the first run.
9:25PM INF No pid for Connectors
9:25PM INF Started process: /home/francd/c8run-8.9.0/connectors.process pid=21641
9:25PM DBG Failed to read PID from file. This is expected for the first run.
9:25PM INF No pid for Camunda
9:25PM INF Started process: /home/francd/c8run-8.9.0/camunda.process pid=21642
9:25PM INF Waiting for service to start name=Camunda retries_left=24
9:25PM INF Waiting for service to start name=Camunda retries_left=23
✅ Camunda has started successfully.
-------------------------------------------
Access each component using the following URLs:
- Operate: http://localhost:8080/operate
- Tasklist: http://localhost:8080/tasklist
- Admin: http://localhost:8080/admin
Login with:
- Username: demo
- Password: demo
API endpoints:
- Orchestration Cluster API: http://localhost:8080/v2/
- Inbound Connectors API: http://localhost:8086/
- Zeebe API (gRPC): http://localhost:26500
- Camunda metrics: http://localhost:9600/actuator/prometheus
MCP servers:
- Orchestration Cluster: http://localhost:8080/mcp/cluster
Note: When using the Desktop Modeler, Authentication may be set to None.
Next steps:
1. Model and deploy a process with Desktop Modeler -> Cluster endpoint: http://localhost:8080/v2/
2. View your process in Operate
3. Run a job worker via Java/Node SDK
4. Build your first AI agent
5. Connect any AI agent to the Orchestration API MCP Server and chat with your processes
Need guidance?
- Quickstart:
https://docs.camunda.io/docs/next/guides/getting-started-example/
- Java developers:
https://docs.camunda.io/docs/guides/getting-started-java-spring/
- Build an agent:
https://docs.camunda.io/docs/next/guides/getting-started-agentic-orchestration/
Run `./c8run help` to see available commands and options.
-------------------------------------------
9:25PM INF All processes are running and healthy, exiting script...