Pull content to files
Schemas and content live in .supatent/schema/ and .supatent/content/, so agents can work in the same file system as the rest of your app.
npx @supatent/cli pullTell us a bit about yourself and your project
Supatent lets agents work the way they already do: edit files on disk. The CLI validates and syncs draft content to the platform, and your team controls which version gets published.
{ "headline": "The CMS built for AI coding agents", "cta": "Request access", "status": "draft"}Keep CMS work where AI coding agents already operate: in your repository. The CLI handles sync and validation; Supatent keeps the draft and published states separate.
Schemas and content live in .supatent/schema/ and .supatent/content/, so agents can work in the same file system as the rest of your app.
npx @supatent/cli pullThe CLI validates local changes and writes machine-readable status for coding agents to inspect and fix.
npx @supatent/cli validate --jsonThe CLI syncs local changes to the remote draft state in Supatent. Agents keep working against draft content.
npx @supatent/cli pushA person reviews the draft in Supatent and promotes it to the immutable published version when it is ready to go live.
Use @supatent/client for published content and assets. The same package includes browser, React, Astro, webhook, markdown asset, and preview-mode helpers.
import { createClient } from "@supatent/client";
const cms = createClient({
baseUrl: "https://app.supatent.ai",
projectSlug: "marketing-site",
locale: "en-US",
});
const page = await cms.schema("landing-page").bySlug("home");
const posts = await cms.schema("blog-post").paginate({ limit: 10 });
const hero = await cms.asset("hero-image");Schema names, locales, assets, and preview behavior all stay in the same client surface instead of being split across bespoke fetch logic.
Supatent handles the serving layer for you. Published content is already cached, invalidated on publish, and exposed through APIs built for production delivery. Images are returned with optimized URLs and responsive variants, so there is no extra optimization pipeline to build before your app consumes the content.
Once a version is published, Supatent makes it available with caching and asset optimization already applied, so your app or website gets production-ready content without an extra delivery pipeline.
The CLI handles local sync. The client handles application reads. Skills and docs tooling give coding agents the project-specific instructions and content surfaces they need to work cleanly.
Sync schemas and content between local files and Supatent. Use init, pull, validate, status, push, dev, and merge in agent workflows.
npx @supatent/cli devQuery published content from TypeScript apps, with subpath exports for browser, React, Astro, and webhook verification.
npm install @supatent/clientInstall Codex and Claude Code skills that teach agents how to manage Supatent schemas, blog posts, landing pages, and SEO-oriented content.
npx @supatent/skillsRender documentation from Supatent content in Next.js. The package supports static-first docs, search assets, and static AI outputs.
npm install @supatent/supatent-docsSupatent treats structured metadata and translations as first-class content work. Model JSON-LD fields, maintain locale-specific files, and let agent skills help with SEO reviews and localized drafts.
schema/ defines the content model. Here, blog-post.json describes the fields that every blog post item must follow.
content/ stores the actual entries. The blog-post/ folder maps directly to the blog-post schema, and each item inside it is a localized version of the same content item.
assets/ keeps the binary file and its locale-aware metadata together. The .png is the image itself, while the adjacent locale JSON files can hold translated fields such as title, alt text, caption, and other per-locale metadata.
Use structured JSON-LD fields for pages, articles, FAQs, software, and other SEO-relevant content models.
Use @supatent/supatent-docs to render documentation from Supatent content in a Next.js app.
Supatent keeps the AI workflow useful without letting it bypass your release process. Drafts can change quickly; published versions are the controlled output your app reads.
CLI and app edits target draft content while the team iterates. This is the working state agents and humans update.
Publishing creates an immutable version for public reads. Your app only consumes the approved output.
A previous published version can be copied back into draft without rewriting the underlying content model.
Schemas, content, assets, and locales stay under one project instead of splitting content governance across multiple tools.
Agents edit Supatent schemas and content as local files. The CLI validates those files and syncs draft changes back to the platform.
No. The agent workflow is draft-first. Publishing remains an explicit action in Supatent so a person controls what goes live.
Use @supatent/client to query published content and assets from your application. Add an API key when you need preview or draft mode during development.
Use @supatent/cli for file sync, @supatent/client for app integration, @supatent/skills for agent workflows, and @supatent/supatent-docs for rendering docs from Supatent content.
Supatent supports JSON-LD fields, multi-locale content, and SEO-focused skills. Add a locale such as fr-FR, ask your coding agent to translate the content files, validate, and push the draft for review.
Start with the CLI, connect your app with the client package, and keep publication under your control.