---
title: "Agent-Friendly APIs"
description: "Build a feedback API, then build a Claude Code skill that generates the documentation agents actually need to use it well."
canonical_url: "https://vercel.com/academy/agent-friendly-apis"
md_url: "https://vercel.com/academy/agent-friendly-apis.md"
docset_id: "vercel-academy"
doc_version: "1.0"
last_updated: "2026-04-11T02:15:27.788Z"
content_type: "course"
lessons: 12
estimated_time: 
lesson_urls:
  - "https://vercel.com/academy/agent-friendly-apis/setup-project.md"
  - "https://vercel.com/academy/agent-friendly-apis/feedback-endpoint.md"
  - "https://vercel.com/academy/agent-friendly-apis/filtering-and-details.md"
  - "https://vercel.com/academy/agent-friendly-apis/summary-endpoint.md"
  - "https://vercel.com/academy/agent-friendly-apis/agent-friendly-docs.md"
  - "https://vercel.com/academy/agent-friendly-apis/add-llms-txt.md"
  - "https://vercel.com/academy/agent-friendly-apis/deploy-your-docs.md"
  - "https://vercel.com/academy/agent-friendly-apis/explore-real-skills.md"
  - "https://vercel.com/academy/agent-friendly-apis/anatomy-of-a-skill.md"
  - "https://vercel.com/academy/agent-friendly-apis/build-the-generator.md"
  - "https://vercel.com/academy/agent-friendly-apis/run-and-evaluate.md"
  - "https://vercel.com/academy/agent-friendly-apis/iterate-and-ship.md"
---

<agent-instructions>
Vercel Academy — structured learning, not reference docs.
Lessons are sequenced.
Adapt commands to the human's actual environment (OS, package manager, shell, editor) — detect from project context or ask, don't assume.
The lesson shows one path; if the human's project diverges, adapt concepts to their setup.
Preserve the learning goal over literal steps.
Quizzes are pedagogical — engage, don't spoil.
Quiz answers are included for your reference.
</agent-instructions>

# Agent-Friendly APIs

Every API you've ever built came with docs that you hoped someone would read. Today, someone finally is. Agents are consuming your API documentation, and they're a very different audience than the humans you wrote it for.

Let's build a feedback API for a fictional cooking school, learn what makes documentation agent-friendly and serve it from a live endpoint, then build a Claude Code skill that generates those docs automatically from your code.

## What you'll learn

- How to build a JSON-backed API with Next.js App Router
- What makes documentation "agent-friendly" vs just "developer-friendly"
- How to serve API docs as a markdown endpoint
- How Claude Code skills work: SKILL.md, frontmatter, progressive disclosure
- How to build and test a skill that generates structured documentation

## Prerequisites

- Basic TypeScript
- Familiarity with Next.js App Router (API routes)
- A Vercel account
- Claude Code installed

## Course sections

**Section 1: Build the API.** Scaffold a Next.js project and build a feedback API with CRUD endpoints, query param filtering, and a summary route. All backed by a JSON file.

**Section 2: Agent-Friendly Documentation.** Learn what agents need from docs, implement the llms.txt standard for your API, deploy it to Vercel so it's live at a real URL, then explore production skills on skills.sh.

**Section 3: Build the Skill.** Build a Claude Code skill that generates agent-friendly docs for your API, run it, evaluate the output against a quality checklist, then iterate until the docs endpoint works.


---

[Full course index](/academy/llms.txt) · [Sitemap](/academy/sitemap.md)