directus
Build backends and APIs with Directus headless CMS. Use when a user asks to create a headless CMS, build a content API without coding, set up a backend admin panel, create REST or GraphQL APIs from a database, manage content with roles and permissions, build a data platform with auto-generated APIs, or replace traditional CMS with a headless solution. Covers data modeling, auto-generated REST/GraphQL APIs, roles/permissions, flows (automation), file storage, and SDK integration.
Usage
Getting Started
- Install the skill using the command above
- Open your AI coding agent (Claude Code, Codex, Gemini CLI, or Cursor)
- Reference the skill in your prompt
- The AI will use the skill's capabilities automatically
Example Prompts
- "Review the open pull requests and summarize what needs attention"
- "Generate a changelog from the last 20 commits on the main branch"
Documentation
Overview
Directus is an open-source headless CMS and data platform that wraps any SQL database with auto-generated REST and GraphQL APIs, a visual admin dashboard, role-based access control, file storage, and automation flows. Unlike Strapi (which defines its own schema), Directus mirrors your existing database — add columns in the admin UI or directly in SQL, and the API updates instantly. Use it for content management, internal tools, data APIs, and backend-as-a-service.
Instructions
Step 1: Deployment
# Docker (quickest start)
docker run -d --name directus \
-p 8055:8055 \
-e SECRET="your-secret-key-min-32-chars" \
-e ADMIN_EMAIL="admin@example.com" \
-e ADMIN_PASSWORD="your-secure-password" \
-e DB_CLIENT="sqlite3" \
-e DB_FILENAME="/directus/database/data.db" \
-v directus_data:/directus/database \
-v directus_uploads:/directus/uploads \
directus/directus:latest
# Production with PostgreSQL
docker run -d --name directus \
-p 8055:8055 \
-e SECRET="your-secret-key" \
-e ADMIN_EMAIL="admin@example.com" \
-e ADMIN_PASSWORD="your-secure-password" \
-e DB_CLIENT="pg" \
-e DB_HOST="postgres" \
-e DB_PORT="5432" \
-e DB_DATABASE="directus" \
-e DB_USER="directus" \
-e DB_PASSWORD="dbpassword" \
directus/directus:latest
# Access admin: http://localhost:8055
Step 2: Data Modeling
Create collections (tables) and fields via the admin UI or REST API:
# Create a "posts" collection via API
curl -X POST http://localhost:8055/collections \
-H 'Authorization: Bearer admin_token' \
-H 'Content-Type: application/json' \
-d '{
"collection": "posts",
"meta": { "icon": "article", "note": "Blog posts" },
"fields": [
{ "field": "id", "type": "uuid", "meta": { "special": ["uuid"] }, "schema": { "is_primary_key": true } },
{ "field": "title", "type": "string", "meta": { "required": true } },
{ "field": "slug", "type": "string", "meta": { "interface": "input" } },
{ "field": "content", "type": "text", "meta": { "interface": "input-rich-text-html" } },
{ "field": "status", "type": "string", "meta": { "interface": "select-dropdown", "options": { "choices": [{"text":"Draft","value":"draft"},{"text":"Published","value":"published"}] } } },
{ "field": "published_at", "type": "timestamp" }
]
}'
Step 3: Auto-Generated APIs
Once collections exist, Directus auto-generates full CRUD APIs:
# REST — List all published posts
curl 'http://localhost:8055/items/posts?filter[status][_eq]=published&sort=-published_at&limit=10' \
-H 'Authorization: Bearer token'
# REST — Get single post with related author
curl 'http://localhost:8055/items/posts/POST_ID?fields=*,author.name,author.avatar' \
-H 'Authorization: Bearer token'
# REST — Create post
curl -X POST http://localhost:8055/items/posts \
-H 'Authorization: Bearer token' \
-H 'Content-Type: application/json' \
-d '{"title": "My Post", "content": "<p>Hello world</p>", "status": "draft"}'
# GraphQL — Same queries
curl -X POST http://localhost:8055/graphql \
-H 'Authorization: Bearer token' \
-H 'Content-Type: application/json' \
-d '{"query": "{ posts(filter: {status: {_eq: \"published\"}}, sort: [\"-published_at\"], limit: 10) { id title content published_at author { name } } }"}'
Step 4: SDK Integration
// lib/directus.js — JavaScript SDK for frontend/backend integration
import { createDirectus, rest, readItems, createItem, authentication } from '@directus/sdk'
const client = createDirectus('http://localhost:8055')
.with(authentication())
.with(rest())
// Fetch published posts
const posts = await client.request(
readItems('posts', {
filter: { status: { _eq: 'published' } },
sort: ['-published_at'],
limit: 10,
fields: ['id', 'title', 'slug', 'content', 'published_at', { author: ['name', 'avatar'] }],
})
)
// Create a new post
const newPost = await client.request(
createItem('posts', {
title: 'New Post',
content: '<p>Content here</p>',
status: 'draft',
})
)
Step 5: Roles and Permissions
# Create a read-only "viewer" role
curl -X POST http://localhost:8055/roles \
-H 'Authorization: Bearer admin_token' \
-H 'Content-Type: application/json' \
-d '{"name": "Viewer", "admin_access": false}'
# Set permissions: viewer can read published posts only
curl -X POST http://localhost:8055/permissions \
-H 'Authorization: Bearer admin_token' \
-H 'Content-Type: application/json' \
-d '{
"role": "VIEWER_ROLE_ID",
"collection": "posts",
"action": "read",
"permissions": { "status": { "_eq": "published" } },
"fields": ["id", "title", "content", "published_at"]
}'
Step 6: Flows (Automation)
Directus Flows are visual automation pipelines triggered by events (like Zapier, but built-in).
# Create a flow: when a post is published, send a webhook
curl -X POST http://localhost:8055/flows \
-H 'Authorization: Bearer admin_token' \
-H 'Content-Type: application/json' \
-d '{
"name": "Notify on Publish",
"trigger": "event",
"options": { "type": "action", "scope": ["items.update"], "collections": ["posts"] },
"accountability": "all",
"status": "active"
}'
Examples
Example 1: Build a content API for a marketing website
User prompt: "I need a CMS backend for our marketing site — blog posts, team members, case studies, and FAQ. Non-technical editors should be able to manage content through a visual dashboard."
The agent will:
- Deploy Directus with Docker + PostgreSQL.
- Create collections for posts, team_members, case_studies, and faqs.
- Set up relational fields (posts → author, case studies → tags).
- Configure a public role with read-only access for the frontend.
- Connect the Next.js/Astro frontend using the Directus SDK.
Example 2: Build an internal tool for operations data
User prompt: "Our ops team tracks orders, suppliers, and inventory in spreadsheets. Build a proper backend with an admin panel where they can manage everything."
The agent will:
- Deploy Directus pointing at the existing PostgreSQL database.
- Directus auto-detects existing tables and generates APIs + admin UI.
- Create roles: admin (full access), manager (CRUD), viewer (read-only).
- Set up flows for notifications (new order → Slack alert).
Guidelines
- Directus mirrors your database schema — it does not own it. You can add columns via Directus admin or directly in SQL, and both stay in sync. This makes it safe for existing databases.
- Use Directus as a backend-as-a-service for content-heavy apps. For complex business logic (multi-step workflows, custom calculations), extend with custom endpoints or use a separate API layer.
- Configure the
PUBLICrole carefully — it defines what unauthenticated users can access. For a public blog, allow read access to published posts only. - File uploads go to local storage by default. In production, configure S3, Cloudflare R2, or Google Cloud Storage for scalability.
- Directus supports PostgreSQL, MySQL, MariaDB, MS SQL, SQLite, CockroachDB, and OracleDB. PostgreSQL is recommended for production.
Information
- Version
- 1.0.0
- Author
- terminal-skills
- Category
- Development
- License
- Apache-2.0