Frequently Asked Questions

Q: How long does architecture generation take? Initial generation takes 30-60 seconds. The Workflow Architect may ask clarifying questions before generating the spec. Typical output is 3-7 agents depending on use case complexity.

Q: Can I modify the generated agents? Yes. Expand any agent card to edit its name, description, model, tools, memory, visibility, system prompt, and agent-as-a-tool settings. You can also refine in plain English using the Change with AI panel.

Q: How many models are available? 24+ models across providers including Anthropic (Claude Opus 5, Sonnet 4.5, Haiku 4.5), Amazon Nova, Meta Llama, Mistral, NVIDIA Nemotron, Qwen, MiniMax, and Z.AI.

Q: What tools can I add to agents? RAG (Knowledge Base), MCP (external integrations), Web Search, Generate Report, Text to Speech, Speech to Text, and Video tools.

Q: How does deployment work? Deployment fires automatically once the Workflow Architect finishes generating the agent spec - there is no manual Approve step. When you're ready to push to production, click the Deploy button in the Agentic Workflows tab and choose between Qubitz Cloud or BYOC (Bring Your Own Cloud).

Q: How do I test my agents? Click Run on any workflow card in the Agentic Workflows section. The Test Bed opens for that workflow - enter inputs, start the run, and watch the live output stream. Run evaluations across up to 8 metrics, and run safety & stress tests with adversarial attack techniques.

Q: How many workflows can I have per project? Each project supports a maximum of 30 workflows, with a total storage cap of 300 KB per project and 24 KB per individual workflow. The UI shows remaining capacity before you reach the limit. If a project is full, you will see a clear message before any save is attempted.

Q: How long does app code generation take? Typically 3-5 minutes. After code generation, the Amplify build takes additional time. A live log stream shows every step as it runs.

Q: How many API keys can I generate? Each project supports a maximum of 2 API keys. Keys are shown only once at generation - store them securely.

Q: What is the App Refiner? After a successful app generation, the App Refiner lets you make AI-assisted changes to your live app by describing what you want in plain English. Changes deploy automatically at the same URL.

Q: What is the Business Case? Available from the App Generator tab in Application Management. It has three sub-tabs: Cost Estimates (itemized AWS service costs across Conservative, Base, and Growth scenarios), AWS Architecture Diagram, and Export as PDF with an AWS Pricing Calculator link.

Q: What is GraphRAG? GraphRAG is the domain knowledge graph builder inside the Data section. It constructs a property knowledge graph from your uploaded documents - nodes, edges, and evidence anchors - visualized on an interactive Cytoscape.js canvas. Agents can query the graph at runtime for structured relationship lookups. See GraphRAG for the full guide.

Q: What AWS services does Control Hub deploy to? Amazon Bedrock AgentCore, ECS Fargate, DynamoDB, S3, Cognito, and CloudWatch. The infrastructure is fully managed by Qubitz (Qubitz Cloud) or deployed into your own AWS account (BYOC).

Q: Can I update my agents after deployment? Yes. Go to the Agentic System section, make changes via the agent cards or the Change with AI panel, and redeploy. Use the Test Bed to validate before generating a new app build.