Advanced

This guide covers advanced Directus features in n8n using raw CRUD operations, which give you full access to Directus's native filter syntax, query parameters, and complex query capabilities.

← Back to Directus + n8n Overview

Raw CRUD Operations

The Directus node provides raw versions of all CRUD operations that allow you to use Directus's native JSON syntax for filters, query parameters, and data manipulation. These operations give you full control over the API request.

Available Raw Operations

Quick reference of all available raw operations organized by resource type:

Using Raw Operations

Raw operations work similarly to their standard counterparts, but instead of using the node's form fields, you provide data as a JSON object: the JSON Data field for Create/Update, or the Query Parameters field for Get/Get Many.

Setting Up a Raw Operation

1. Add a Directus node to your workflow
2. Set Resource to Item, User, or File
3. Select a Raw JSON operation (e.g., "Get Many (Raw JSON)", "Create (Raw JSON)")
4. For Items operations, select the Collection
5. Enter your data in the JSON Data field (Create/Update) or the Query Parameters field (Get/Get Many)
6. For Get and Update operations, provide the Item ID if needed

Get Many (Raw JSON)

Retrieve items with advanced filtering and query parameters. Get Many (Raw JSON) returns one page of results from the query JSON as written.

{
  "filter": {
    "status": {"_eq": "published"},
    "views": {"_gt": 1000}
  },
  "fields": "title,author.name,views",
  "sort": "-date_created",
  "limit": 10
}

limit controls that single request:

• Omit limit to get Directus's default page (usually 100), not the whole collection.
• Set "limit": 10 to get 10 rows.
• Set "limit": -1 to request Directus's maximum allowed page size. If QUERY_LIMIT_MAX is set (for example, 5), you still only get that many rows. This is not a dump of the whole collection.

To process a large filtered set in chunks, keep limit and offset in the query and use n8n's Split In Batches node.

If you want all results and you don't need a custom filter, use standard Get Many with Return All.

Get (Raw JSON)

Retrieve a single item with custom query parameters:

{
  "fields": "title,content,author.name,author.email,comments.text"
}

Create (Raw JSON)

Create items with full control over the JSON structure:

{
  "title": "My New Post",
  "content": "Post content here",
  "status": "published",
  "author": "author-uuid-here",
  "categories": ["category-uuid-1", "category-uuid-2"]
}

Update (Raw JSON)

Update items with complex data structures:

{
  "title": "Updated Title",
  "status": "archived",
  "metadata": {
    "tags": ["updated", "archived"],
    "notes": "Item has been archived"
  }
}

Using Filters with Raw Operations

Raw operations allow you to use Directus's complete filter syntax. Specify filters in the filter parameter of the Query Parameters field (Get Many).

Example: Complex filter with logical operators

{
  "filter": {
    "_and": [
      {"status": {"_eq": "published"}},
      {
        "_or": [
          {"category": {"_eq": "tutorial"}},
          {"category": {"_eq": "guide"}}
        ]
      },
      {"views": {"_gt": 100}}
    ]
  }
}

Example: Relational filtering

{
  "filter": {
    "author": {
      "name": {"_eq": "John Doe"}
    },
    "comments": {
      "_some": {
        "status": {"_eq": "approved"}
      }
    }
  }
}

Query Parameters

Get (Raw JSON) and Get Many (Raw JSON) support all Directus query parameters. Include them in the Query Parameters field alongside filters. If you omit limit on Get Many (Raw JSON), Directus returns its default page (usually 100), not the whole collection.

Common query parameters:

{
  "fields": "title,author.name,views",
  "sort": "-date_created",
  "limit": 10,
  "offset": 0
}

Search:

{
  "search": "directus tutorial"
}

Aggregation:

{
  "aggregate": {
    "count": "id",
    "sum": "views",
    "avg": "rating"
  },
  "groupBy": "category"
}

Working with Relations

When using Create (Raw JSON) or Update (Raw JSON), you can include related data directly in your JSON:

{
  "title": "My Post",
  "author": "author-uuid-here",
  "categories": ["category-uuid-1", "category-uuid-2"],
  "comments": [
    {"text": "Great post!", "user": "user-uuid-here"}
  ]
}

Using Expressions in Raw Operations

You can use n8n expressions in your raw JSON data for dynamic queries:

{
  "filter": {
    "status": {"_eq": "{{ $json.status }}"},
    "views": {"_gt": {{ $json.min_views }}}
  },
  "fields": "{{ $json.requested_fields }}",
  "sort": "-{{ $json.sort_field }}",
  "limit": {{ $json.limit }}
}

Performance Tips

• Select only the fields you need: Use the fields parameter to reduce data transfer
• Filter in Directus: Use the filter parameter rather than loading extra rows into n8n
• Paginate in chunks: Use limit and offset for large datasets, then process batches with n8n's Split In Batches node. Get Many (Raw JSON) returns one page of results.
• Return All is all-at-once: Standard Get Many with Return All returns all results into n8n in one node run. Use it when you need the rest of a set and you don't need a custom filter. Do not use it to dump a large collection, or when you need to write out rows in chunks.

Example:

{
  "fields": "id,title,status",
  "filter": {
    "status": {"_eq": "published"},
    "date_created": {"_gte": "$NOW(-30 days)"}
  },
  "limit": 100
}

Next Steps

• ← Back to Overview Return to the integration overview
• Learn about Directus Actions → Standard Get Many, Return All, and other operations
• Learn about Directus Triggers → Automation workflows