Content infrastructure for modern teams.
Model your content once. Deliver it anywhere — through REST, GraphQL, or an MCP server your AI tools can speak to directly.
Deliver anywhere
One model. Three doors.
Every project ships with three delivery channels, all reading from the same content — pick per consumer, not per project.
REST API
Predictable resources, cursor pagination, conditional requests. The workhorse.
GraphQL
One query, exactly the shape your front-end needs. Full schema introspection.
MCP server
28 tools your AI assistant calls directly — read, write, publish, model. OAuth-scoped.
Enable Content Teams
Edit content, not code.
Create a content type, add your fields, and it's live over the API the same second — no migrations, no deploys, nothing to write. Powerful enough for the most technical setups, simple enough that anyone can model real data and ship it.
Essay
The quiet web is bigger than you think
There is a kind of internet that does not optimise for you. It does not collect metrics. It refuses to recommend…
{
"mcpServers": {
"headless": {
"url": "https://headless.build/api/mcp",
}
}
}Built for AI assistants
Give your AI the keys.
Allow your AI assistant to search semantically, draft entries, manage media and publish — with exactly the permissions you grant, nothing more.
When something changes
Keep your tools in the know.
Subscribe to any of 11 events and Headless will POST a signed payload the moment it fires — to Slack, to an n8n workflow, or straight to your own service.
Slack
Post to a channel the moment an entry publishes — your team sees it before the cache warms.
n8n
Trigger a flow on any event — fan out to a CDN purge, a sitemap rebuild, a digest.
Your endpoint
Any HTTPS URL. Signed, retried up to 6 times, every attempt logged. SSRF-guarded.