When nothing in the catalog fits, define your own tool once and give an agent a real, reusable capability instead of hardcoding logic into a single flow.
What it can do
- A named tool with a real description — the agent decides when to call it the same way it decides between any built-in tool.
- An input schema — define exactly what arguments the tool expects.
- A JavaScript function body — your own logic, running server-side.
- Reusable across flows — pick it from a dropdown anywhere a Tool node is used, instead of redefining it per flow.
- Return Direct — optionally skip the agent's own summarizing pass and send the tool's output straight back to the user.
Good to know
This is the building block behind anything Flowera doesn't ship natively — in practice, the same escape hatch as HTTP Requests, but with real code instead of just a request shape.