Build Agent Workflows
Build your own AI agents that call APIs as tools: a research agent that cites its sources, an agent that reviews suspicious signups, and the tests and safeguards both need before launch.
Build Agent Workflows badgeComplete all 4 lessons to earn it. About 3 hours.
0 of 4 lessonsStart course
What you'll be able to do
- Generate tool definitions from an API's published parameters and run the tool-use loop server-side
- Build a research agent that reads pages, summarises them and cites only sources it actually read
- Build a triage agent that gathers verification evidence and writes a recommendation for a human
- Test agents against a fixed case set, checking which tools were called and with what arguments
- Add budgets, input validation, human approval and structured logs before an agent goes live
This course is one of three skill areas on exam APV-220, APIVerve Certified Developer: AI Agents: 10 of its 30 questions. See the full path
Syllabus
- Use REST APIs as LLM tools in your own codeStart hereTurn REST endpoints into LLM tools: build JSON schemas from API parameters, run the tool-use loop and execute every API call server-side in Node and Python.
- Build a research agent that cites its sourcesBuild a research agent that reads web pages with a URL to Markdown API, summarises them and cites only sources it read, with quotes checked in code.
- Build an agent that reviews suspicious signupsBuild an agent that reviews flagged signups with email validation, VPN detection and IP lookup APIs, then writes an evidence-backed recommendation for a human.
- Test AI agents and add guardrails before launchTest AI agents with a fixed case set and recorded API responses, check tool-call correctness, and add budgets, approval gates and redacted logs.