Get started
Make your first edit (5 minutes)
Boot the stack locally, point at the demo site, type a chat message, watch it apply in real time. → QuickstartUnderstand the model before writing any code
Pages, blocks, props, operations, sessions, snapshots. The mental model that everything else builds on. → Core Concepts · How It Works · ArchitectureBring in your own site
Migrate an existing Next.js site (agentic, 5–15 min)
Give the in-editor Onboarding agent a URL or GitHub repo. It clones, analyzes, wires the SDK, and registers the site. → Onboarding agentHand the integration to your coding agent (Claude Code, Cursor, Codex)
Self-contained docs the coding agent can read and execute against without supervision. → Bring your own coding agentIntegrate by hand (~30 min)
Drop the Site SDK into your Next.js 15 project manually. Best when you want to understand every wire. → Manual integration · Next.js integration referenceConnect a CMS
Wire up Contentful, Sanity, or Strapi
All three ship as working examples underexamples/ with bootstrap scripts that generate the content model for you.
→ CMS adapters
Write an adapter for a CMS we don’t ship (Storyblok, Hygraph, Payload, Directus, …)
Two functions:fetchPage(slug) → PageDoc and writePage(doc) → void. ~150 lines.
→ CMS adapters — Writing your own
Customize what’s editable
Add a custom block alongside the 20 built-ins
Register a React component with a Zod schema and field metadata. The AI planner reads the schema and edits it like any built-in block. → Custom blocks · Block systemBrowse the live catalogue of built-in blocks
20 blocks, each with editable props, in a live workspace with viewport switcher. → avocadostudio.dev/components · Built-in blocks referenceUse drag-and-drop instead of (or alongside) chat
Avocado ships a Puck integration that produces the sameBlockInstance model from a visual editor.
→ Puck mode
Drive Avocado from outside the editor
From Claude Desktop, Claude Code, or Cursor
Avocado bundles an MCP server with 40 tools. Drop in a config snippet and your AI assistant can read and edit your site directly. → MCP serverFrom a Jira ticket
Webhook + REST integration that turns ticket comments into chat messages. → Jira integrationDeploy
Self-host with Docker
The supported production path. Docker image + persistent volume + Render / Fly / Railway / DO / Kubernetes. → Docker deploymentDeploy the editor + site to Vercel
Three projects: orchestrator (Docker/Render), editor (Vercel), site (Vercel). → Vercel deployment · Netlify deploymentRun a public playground / demo
Locked-downDEMO_MODE=1 with allow-listed ops, per-IP rate limiting, and no AI image gen.
→ Demo mode
Publish to a custom target (S3, GitLab Pages, …)
ImplementPublishTarget (two methods) and register it. The route handler picks it up automatically.
→ Publishing — Building a custom target
Tune AI behavior
Use Gemini (or OpenAI, or both) instead of Claude
Set the appropriate API key. The Content Studio’s model picker surfaces any provider with a configured key. → AI Providers & Model RoutingUse cheaper models for routine edits, smarter ones for restructures
Per-tier model env vars (*_MODEL_FAST, *_MODEL_BALANCED, *_MODEL_REASONING, *_MODEL_CODEX).
→ AI Providers — Model tiers
Add a new language to the editor + AI responses
One new dictionary file, three lines of glue, one entry in the orchestrator’sLOCALE_NAMES.
→ Internationalization