Agent Studio
Chain prompts and AI steps into automated pipelines. Each step's output feeds the next — build multi-step workflows that run with one click.
Agent Details
Governance & Approval
Controls how autonomously this agent can act
Pipeline Steps
Each step receives the previous step's output as {{prev}}
Prompt
Send a custom prompt to the LLM
Summarise
Shorten the previous output
Extract
Pull out specific fields as JSON
Classify
Assign a category label
RAG Query
Search knowledge base and inject context
Format
Reformat output using a template
Run Agent
Waiting for Approval
This agent requires human sign-off before the output is finalised. A pending approval has been added to the queue above.
Recent Runs
No runs yet