Agent Component

Where we left off

Lesson 13 gave an Agent a custom tool you wrote yourself. Most of the time you don't need to, Langflow's own Agent component ships two tools as checkboxes, a calculator and a current-date lookup, no code at all. This lesson is the plain version: same Agent, same wiring, no Custom Component in sight.

Do this yourself

1. Build Chat Input, Google Generative AI, and Agent, wire Google Generative AI's Model output into the Agent's Model input (Lesson 13's connection, a model-typed port, not the message one), Chat Input into the Agent's Input.
2. Open the Agent's settings. Find Current Date and Calculator toggles (under the tools section) and turn both on. No Custom Component, no wiring, a checkbox each.
3. Wire the Agent's Response into a Chat Output.

No canvas.png ships with this lesson: re-importing a saved flow that wires a model into an Agent's Model field this way trips a frontend re-import quirk in this Langflow version, the edge gets silently dropped with an "invalid connections" notice, even though the flow runs correctly (verified via lesson.py). Build it by hand following the steps above and the wiring holds.

1. In the Playground, ask something that needs both: "what is 47 times 6, and what's today's date?" Confirm both parts of the answer are correct, the date one especially, since that's not something the model would know on its own.

The code, piece by piece

agent.set(
    model=gemini.build_model,
    add_calculator_tool=True,
    add_current_date_tool=True,
    ...
)

add_calculator_tool and add_current_date_tool are plain boolean fields on AgentComponent, the code-equivalent of the two toggles from step 2. No tools=[...] list needed for these, since they're not separate components at all, just built into the Agent itself.

result = run_flow_from_json(flow=str(FLOW_PATH), input_value="What is 47 times 6, and what's today's date?")

Back to run_flow_from_json, same as every lesson before 12-13. Unlike those two, nothing here involves a Custom Component, so there's no JSON-deserialization risk to route around, this flow loads and runs completely normally.