The Solution Suite is the set of surfaces you use to turn an idea into a governed, observable AI system running in your business. Vibe Studio for scaffolding, the Agent Builder for hand-tuning, Workflows for orchestrating multiple agents, Approvals for the human-in-the-loop moments, and the Runs observability stack for seeing everything that happened.
The pieces
- Vibe Studio - prompt-driven scaffolding for agents, apps, and workflows (see the Vibe Studio guide)
- Agent Builder - the classic hand-tuned agent view when you want fine control over model, tools, prompt, and guardrails
- Workflows - branching pipelines that chain agents with rules, integrations, and human approvals for consequential decisions
- HITL Reviews - the queue where humans approve, reject, or amend actions the agent proposes at checkpoints
- Prompt Lab - a scratchpad for iterating on prompts before you invest in a full agent build
- Runs - the observability tab where every execution logs its cost, latency, tools called, and reasoning trace
How they fit together
A typical flow: you start in Vibe Studio to sketch the idea, refine the specifics in the Agent Builder, wire it into a Workflow that fires on the right event, set the HITL checkpoint for anything consequential, and watch the Runs tab to confirm behaviour matches expectations. Nothing about the flow is prescribed - some teams live entirely in Vibe Studio; others prefer the Builder's finer control.
Everything is versioned + audited
Every published agent snapshots to an immutable version. Every run writes to an audit log with actor, inputs, outputs, and AI Credits consumed. Every HITL decision writes who approved (or rejected) with a reason. This isn't optional or configurable - it's the default because governance-first AI is the platform's whole promise.