Stitch Sdk Readme

Stitch SDK README Generator

This skill produces the README for @google/stitch-sdk. It combines a structural strategy (the Bookstore Test) with instructions for sourcing the current API from the codebase — so the README stays accurate as the SDK evolves.

---

How to Source the Current API

Do not hard-code the API surface. Read it from the codebase at invocation time:

After reading these files, you have the complete API surface. Structure it using the Bookstore Test template below.

---

The Bookstore Test

A reader decides whether to use a library the same way a person decides to buy a book: they glance at the cover, read the inner flap, then commit to reading the book. The README must earn the reader's attention at each stage.

The Cover

A single sentence stating what problem this library solves — not what the library _is_. The reader should recognize their own situation. No taglines, no badges, no logos.

For this SDK, the cover is about generating UI from text and extracting HTML/screenshots programmatically.

Good: "Generate UI screens from text prompts and extract their HTML and screenshots programmatically." Bad: "The official TypeScript SDK for Google Stitch, a powerful AI-powered UI generation platform."

The Inner Flap

Immediately show the library in use. Code first, not setup.

Primary workflow — the punchline everything in the SDK exists to produce:

project(id) → generate → getHtml

Show this as the first code block, with one line noting the env var requirement. Do not show installation, imports, or config before this. Show callTool("create_project", ...) separately for project creation.

Secondary workflows — reveal depth progressively:

1. Listing and iterating over existing projects/screens 2. Editing a screen and generating variants 3. Tool access via singleton (stitch.listTools(), stitch.callTool()) — zero setup 4. Explicit configuration via StitchToolClient (custom API key, base URL) 5. AI SDK integration via stitchTools() — import from @google/stitch-sdk/ai, show generateText with tools: stitchTools() and stepCountIs

Rules for this section:

• No setup first. One line mentioning STITCH_API_KEY is enough before the first example.
• Dual install paths. Show npm install @google/stitch-sdk first (core SDK, standalone). Then show npm install @google/stitch-sdk ai for AI SDK users. The ai package is only needed when importing from @google/stitch-sdk/ai.
• Straightforward language. No "powerful", "seamless", "robust", "enterprise-grade".
• Working examples. Every code block must be valid, runnable code — not fragments with // ... elisions.
• Progressive complexity. Simplest invocation first, then deeper capabilities.

Reading the Book

The reader is committed. Document the full API as a reference.

Structure by class in this order: Stitch → Project → DesignSystem → Screen → StitchToolClient → toolDefinitions / toolMap → stitchTools() (AI SDK) → StitchProxy → stitch singleton.

Each entry should have:

• What it does (one line)
• Usage example (minimal, runnable)
• Parameters (table)
• Return type and error behavior

Setup, authentication, and configuration go here — after the reader has already decided the library is worth using.

Tone

Write like a colleague explaining their work to another engineer. Be direct. Be specific. Don't sell — inform. If a feature has limitations, state them. If setup is complex, say so.

---

Validation

After generating the README, verify:

• [ ] Can a reader understand what the library does in under 10 seconds?
• [ ] Is there a runnable code example within the first scroll?
• [ ] Does setup/config appear _after_ the first code example?
• [ ] Is every code block valid, copy-pasteable code?
• [ ] Is the language descriptive rather than promotional?
• [ ] Does the reference section cover every public export from index.ts?
• [ ] Every method name in examples exists in its class source file
• [ ] Every import in examples matches an export in index.ts
• [ ] All three modalities are documented: domain classes (scripts), StitchToolClient (agents), stitchTools() (AI SDK)

Anti-patterns