Modules 07–08. Prepared n8n and a supported tool-calling chat model.
02
Open these tools
AI Agent, supported Chat Model and Calculator sub-node.
03
Check the requirement
The chosen hosted model may need API billing. A chat subscription is not an API credential.
Module 10 / Build together
Your first small win.
Work out the fictional supplies total: 12 notebooks at 3 units each, plus 8 pens at 2 units each. Show the arithmetic and state the total. This is a workshop calculation, not a real purchase.
Module 10 / Build together
What happens between input and output?
The agent receives a goal and tool descriptions.
The model chooses a permitted tool when needed.
You inspect tool calls and the final answer.
Module 10 / Build together
Step 01. Confirm a supported model.
1Use the facilitator-approved tool-calling model and credentials.
2Check current n8n supported-model documentation.
3Do not assume every Ollama/chat model supports this agent.
Module 10 / Build together
Step 02. Create the input.
1Add Manual Trigger and Edit Fields.
2Create a String question with the supplies calculation.
3Execute the input node and check its value.
Module 10 / Build together
Step 03. Add AI Agent.
1Connect AI Agent and set the prompt from question.
2Attach the supported Chat Model.
3Do not use an obsolete agent-type selector.
Module 10 / Build together
Step 04. Give it one tool.
1Attach Calculator to the agent tool connector.
2Set the system instruction to the scoped agent prompt.
3Set Max Iterations to 3 for this small test.
Module 10 / Build together
Step 05. Inspect execution.
1Run once and inspect the tool call in execution details.
2Confirm 12*3 + 8*2 gives 52.
3Do not assume a plausible final answer proves the tool ran.
Module 10 / Build together
Step 06. Test a boundary.
1Ask it to place the order.
2It should say no purchasing tool is available.
3Keep the tool list read-only for this lesson.
Module 10 / Build together
Make the working version.
You are a workshop supplies assistant. Use Calculator for arithmetic when appropriate. Show the expression and result. Use only supplied quantities. No purchasing, messaging or booking tools are available. Ask for missing quantities; never claim actions were completed.
Module 10 / Build together
Ask for a review. Then check it.
Inspect this recorded agent output and tool trace. Did a calculator call occur? Does the result match 12*3+8*2=52? Flag unsupported claims. Do not assume execution from answer text alone. Trace: [paste redacted trace].
Module 10 / Build together
Prove your result works.
1A real Calculator call is visible.
2The result is 52 for the fictional example.
3The agent does not claim to buy anything.
Module 10 / Build together
When it does not work yet.
01
Tool-call error
Choose a currently supported tool-calling model.
02
Only an answer, no call
Inspect trace and tool connection; adjust the instruction then rerun.
Module 10 / Build together
Choose your next challenge.
01
Core complete
Show one checked output to a partner.
02
Extension 1
Add one bounded read-only lookup tool.
03
Extension 2
Compare agent tool choice with a fixed workflow call.
Module 10 / Build together
Keep the result. Keep the evidence.
Watch a tool-using agent calculate a fictional supplies total.
Save your filesSave your promptName the next change
Module 10 / Build together
Sources you can check.
01
n8n
AI Agent
02
n8n
Supported tool-calling models
03
n8n
Calculator tool
Watch an agent hand off the work.
Original ten-second Hyperframes render. Follow the request through an agent and a tool, then stop for human review. The lesson diagrams above are crisp vector animations; this is the actual rendered video.
Outcome: Watch a tool-using agent calculate a fictional supplies total. This is a 60-minute in-person laptop session after preflight. No setup-duration guarantee. Core first, extensions second.
References
Before you start.
3 minutes
The chosen hosted model may need API billing. A chat subscription is not an API credential. The facilitator verifies accounts, installed tools and model/host connections before the session. Pair up if setup is incomplete. Do not create accounts or incur costs without understanding the provider terms.
References
Your first small win.
5 minutes
Quick win: Inspect a prepared calculator tool call and check its result. Aim for a visible result in the first 15 minutes after preflight. Explain that generated code/output is untested until the learner checks it. If tools are unavailable, inspect the prepared example with a partner.
References
What happens between input and output?
3 minutes
Ask a learner to explain each stage in their own words. This is a teaching model, not a promise that all tools behave identically.
References
Step 01. Confirm a supported model.
3 minutes
Perform these numbered actions on the laptop. Pause after the step and inspect the result. Keep credentials private and use the prepared workshop instance.
Perform these numbered actions on the laptop. Pause after the step and inspect the result. Keep credentials private and use the prepared workshop instance.
Perform these numbered actions on the laptop. Pause after the step and inspect the result. Keep credentials private and use the prepared workshop instance.
Perform these numbered actions on the laptop. Pause after the step and inspect the result. Keep credentials private and use the prepared workshop instance.
Perform these numbered actions on the laptop. Pause after the step and inspect the result. Keep credentials private and use the prepared workshop instance.
Perform these numbered actions on the laptop. Pause after the step and inspect the result. Keep credentials private and use the prepared workshop instance.
Use fictional practice inputs. Ask learners to show their work, not just the final answer.
References
Ask for a review. Then check it.
4 minutes
Use fictional practice inputs. Ask learners to show their work, not just the final answer.
References
Prove your result works.
6 minutes
These are manual acceptance tests for the learner to perform. No test is claimed to have run merely because it appears here. Ask pairs to demonstrate one pass and one corrected failure.
References
When it does not work yet.
4 minutes
Choose the observed symptom. Change one thing and rerun the smallest test. Avoid restarting the whole project or adding tools before diagnosing the current step.
References
Choose your next challenge.
5 minutes
Experienced learners use the extensions while beginners complete the same core outcome. An extension is optional, not a prerequisite for participating.
References
Keep the result. Keep the evidence.
3 minutes
Save the working version before leaving. Add a short note: what worked, what needed correction and what you would try next. The next module builds on the saved output.
References
Sources you can check.
2 minutes
Primary sources checked during course creation on 2026-10-08. Provider interfaces and availability can change. The examples, prompts and activities are original community teaching designs. Source links explain capabilities; they do not prove these fictional exercises were executed.
