Model, tools and memory
Choose its model, attach the tools it can call and add memory when a conversation needs to continue. You define which resources it can use.
Your tools.Your rules.
Orchestrate APIs, data processing and AI agents in one environment. Compose workflows visually, extend them with code and inspect every execution.
The nodulz application, at a glance.
Visual orchestration. Explicit logic.
Build workflows from triggers, conditions and sub-workflows. The graph exposes dependencies and branches; each node retains its configuration.
Triggers, branches and sub-workflows
Webhook, schedule, Git event or manual execution. Connect processing steps, define conditions and reuse logic through sub-workflows.
Your APIs. Your code. One workflow.
Combine HTTP requests, database access and JavaScript or Python transformations. Your scripts can also come from a Git repository, be encapsulated in a container and run on the host, a remote Docker server or in Kubernetes. Their orchestration stays in the same workflow.
HTTPJavaScriptPostgreSQL
normalize-records.jsJavaScript
const records = input.map(record => ({
...record,
status: record.status.toLowerCase()
}));
return records;Input
[{ "status": "ACTIVE" }]Output
[{ "status": "active" }]AI agents, part of the graph.
Attach a model, tools and memory to an agent. Compose RAG pipelines with document loading, chunking and vector search.
document-agentAgent and connected resources
Request
AI Agent
Response
Language model
HTTP / SQL tools
Conversation memory
Vector search
RAG pipelines
Load and chunk documents, then query a vector store to inject relevant passages into the agent context.
Inspect data. Isolate failures.
Execute a node, a chain or the entire workflow. Analyze inputs, outputs and logs in execution history, then configure retry policies.
customer-routingFailed
Notification
InputJSON
{
"destination": "support",
"priority": "high"
}Logs
Invalid notification destination.
Targeted tests
Execute a node with test input, a node chain or the full graph. Validate transformations and connections before a complete execution.
Execution history
Inspect the path, results and logs of each node. Locate a failure and analyze the data received by the affected step.
Error policies
Configure retries and failure behavior: stop, continue or handle the error within the workflow.
Provision services. Connect workflows.
Deploy databases, caches and vector stores from the catalog. Configure instances on Docker or Kubernetes and use them in compatible nodes.
- PostgreSQL
- MariaDB
- MongoDB
- Redis
- Qdrant
Service instances
customer-db
PostgreSQL · Docker
workflow-cache
Redis · Docker
knowledge-store
Qdrant · Kubernetes
Docker and Kubernetes provisioning
Choose a database, cache or vector database, then a Docker or Kubernetes target. Configure your instance and its credentials from nodulz.
Configuration, status and logs
Check instance status and logs. Adjust configuration and trigger a redeployment from the same workspace.
Workflow connections
Use your service instances in compatible nodes. Database and vector database connections integrate with your processing steps and AI agents.
A shared environment. Controlled access.
Centralize workflows and services. Share resources with your team and control access through roles and permissions.
Reusable workflows
Factor out shared processing and call a workflow from multiple automations.
Roles and permissions
Manage workflow and service access through resource roles and permissions.
Documentation and technical reference
Guides and technical reference: coming soon at nodulz.com/docs.
Distribution coming soon.
Move forward at your own pace.
From your first workflows to team collaboration, explore nodulz plans and licensing options.