Documents API

Learn how to add, retrieve, and manage documents in your VectorForgeAI collections.

Document Overview

Documents are the core data units in VectorForgeAI. Each document contains text content that is automatically processed, chunked, and embedded for vector search and retrieval. The Documents API allows you to manage these documents within your collections.

List Documents in a Collection

Retrieve all documents within a specific collection.

Path Parameters

Query Parameters

Request

curl -X GET "https://api.vectorforgeai.com/v1/collections/abc123xyz456/documents?limit=20&page=1" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Team-Token: YOUR_TEAM_TOKEN"

Response

{
  "documents": {
    "current_page": 1,
    "data": [
      {
        "identifier": "doc-001",
        "title": "What is VectorForgeAI",
        "body": "VectorForgeAI is a platform that simplifies AI infrastructure...",
        "metadata": {
          "author": "Jane Smith",
          "department": "Marketing"
        },
        "structured_problems": [],
        "created_at": "2025-05-01T09:15:22.123Z",
        "updated_at": "2025-05-01T09:15:22.123Z"
      },
      {
        "identifier": "doc-002",
        "title": "Getting Started Guide",
        "body": "This guide will help you get started with our platform...",
        "metadata": {
          "category": "Tutorials",
          "difficulty": "Beginner"
        },
        "structured_problems": [{"tags": ["guide", "getting-started"], "problem": "How do I get started with VectorForgeAI?", "summary": "This is a guide how to get started with VectorForgeAI", "solution": "First you go to the signup page, step 2 will be creating an API token. After this you can follow the documentation step by step."}],
        "created_at": "2025-05-02T14:25:12.789Z",
        "updated_at": "2025-05-02T14:25:12.789Z"
      }
    ],
    "first_page_url": "https://api.vectorforgeai.com/v1/collections/abc123xyz456/documents?page=1",
    "from": 1,
    "last_page": 3,
    "per_page": 20,
    "to": 20,
    "total": 42
  }
}

Add a Document to a Collection

Add a new document to an existing collection. If a document with the same identifier already exists, it will be updated.

Path Parameters

Request Parameters

Request

curl -X POST https://api.vectorforgeai.com/v1/collections/abc123xyz456/documents \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Team-Token: YOUR_TEAM_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "identifier": "pricing-guide-2025",
    "title": "VectorForgeAI Pricing Guide 2025",
    "body": "VectorForgeAI offers flexible pricing plans to meet the needs of individuals, startups, and enterprises...",
    "metadata": {
      "category": "Documentation",
      "updated_date": "2025-05-01",
      "version": "2.4"
    }
  }'

Response

{
  "state": "created",
  "document": {
    "identifier": "pricing-guide-2025",
    "title": "VectorForgeAI Pricing Guide 2025",
    "body": "VectorForgeAI offers flexible pricing plans to meet the needs of individuals, startups, and enterprises...",
    "metadata": {
      "category": "Documentation",
      "updated_date": "2025-05-01",
      "version": "2.4"
    },
    "structured_problems": [{"tags": ["guide", "getting-started"], "problem": "How do I get started with VectorForgeAI?", "summary": "This is a guide how to get started with VectorForgeAI", "solution": "First you go to the signup page, step 2 will be creating an API token. After this you can follow the documentation step by step."}],
    "created_at": "2025-05-10T10:15:22.456Z",
    "updated_at": "2025-05-10T10:15:22.456Z"
  }
}

Delete a Document

Permanently remove a document from a collection, including all its vector embeddings.

Path Parameters

Request

curl -X DELETE https://api.vectorforgeai.com/v1/collections/abc123xyz456/documents/pricing-guide-2025 \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Team-Token: YOUR_TEAM_TOKEN"

Response

{
  "message": "Document deleted successfully."
}

Document Processing

When you add a document, two things will happen, VectorForgeAI automatically:

Creates Problem-Solution pairs New

1. Turn the content into Markdown
2. Using LLMs, generate one or more summaries, tags, problems and solutions from the content
3. Generates vector embeddings for each of the problem-solution pairs using our embedding model
4. Chunks the document into smaller segments for optimal processing (100-200 words per chunk)
5. Generates vector embeddings for each chunk using our embedding model
6. Stores both the document metadata and embeddings for efficient retrieval
7. Makes the document available for semantic search and AI-powered responses

Creates Chunks

1. Turn the content into Markdown
2. Chunks the document into smaller segments for optimal processing (100-200 words per chunk)
3. Generates vector embeddings for each chunk using our embedding model
4. Stores both the document metadata and embeddings for efficient retrieval
5. Makes the document available for semantic search and AI-powered responses

Best Practices

• Document Size: While there's no hard limit on document body size, we recommend keeping documents focused on specific topics for better search relevance. We automatically convert HTML to Markdown.
• Identifiers: Use consistent, meaningful identifiers to make document management easier. Can also be the URL.
• Metadata: Use metadata to add searchable attributes to your documents, such as authors, categories, and dates.
• Organization: Group related documents in the same collection for better context during search and retrieval.

Next Steps

Now that you've learned how to manage documents, explore:

• Vector Search - Learn how to search your documents semantically
• LLM Completions - Use your documents as context for AI responses