Build Your Knowledge Base with Documents

What Are Documents?

Documents are files you upload to your Parallel AI workspace to build a knowledge base for your agents. When an agent has knowledgeBaseEnabled: true, it can search through your documents in real time and use the retrieved content to answer user questions with accuracy and context. Use your knowledge base to give agents access to:

Product documentation and FAQs
Internal policies and procedures
Support runbooks and troubleshooting guides
Sales materials and pricing sheets
Any other reference content your agents should know about

Documents are processed and indexed automatically after upload. Once indexed, your agents can search and retrieve relevant passages in real time during conversations.

How Agents Use Documents

When a conversation requires a factual answer, an agent with knowledgeBaseEnabled: true performs a semantic search against your document index and injects the most relevant passages into its context before responding. You can control which documents an agent can access using knowledgeBaseScope:

`knowledgeBaseScope` value	Behavior
`/`	Agent can search all documents in your workspace
`/shared/Support`	Agent only searches documents in the `/shared/Support` folder
`/products/widget`	Agent only searches documents in the `/products/widget` subfolder

This lets you run multiple specialized agents that each have access to only the documents relevant to their role.

Organize your documents into folders that mirror your agent structure. For example, keep /support docs for your support agent and /sales docs for your sales agent.

Core Operations

Operation	Method	Endpoint
List documents	`GET`	`/api/v0/documents`
Upload a document	`POST`	`/api/v0/documents`
Get a document	`GET`	`/api/v0/documents/{id}`
Delete a document	`DELETE`	`/api/v0/documents/{id}`

Uploading a Document

Prepare your file

Supported formats typically include PDF, plain text, Markdown, and DOCX. Check that your document is well-structured for best retrieval results.

Upload via API

Send a POST /api/v0/documents request with your file and optional metadata such as folder path.

Wait for indexing

The document is processed and indexed automatically. Once complete, agents with the matching knowledgeBaseScope can search it immediately.

curl -X POST "https://api.parallellabs.app/api/v0/documents" \
  -H "X-API-Key: YOUR_API_KEY" \
  -F "file=@/path/to/your-document.pdf" \
  -F "name=Product FAQ" \
  -F "folder=/shared/Support"

import requests

with open("/path/to/your-document.pdf", "rb") as f:
    response = requests.post(
        "https://api.parallellabs.app/api/v0/documents",
        headers={"X-API-Key": "YOUR_API_KEY"},
        files={"file": ("your-document.pdf", f, "application/pdf")},
        data={
            "name": "Product FAQ",
            "folder": "/shared/Support"
        }
    )
print(response.json())

Listing Documents with Pagination

Retrieve all documents in your workspace with optional pagination:

curl "https://api.parallellabs.app/api/v0/documents?page=1&pageSize=20" \
  -H "X-API-Key: YOUR_API_KEY"

Parameter	Type	Description
`page`	integer	Page number, starting from `1`
`pageSize`	integer	Number of results per page

Retrieving a Specific Document

curl "https://api.parallellabs.app/api/v0/documents/{documentId}" \
  -H "X-API-Key: YOUR_API_KEY"

Deleting a Document

curl -X DELETE "https://api.parallellabs.app/api/v0/documents/{documentId}" \
  -H "X-API-Key: YOUR_API_KEY"

Deleting a document removes it from the index. Agents will no longer be able to retrieve content from it. This action is irreversible.

Enabling Knowledge Base on an Agent

Once your documents are uploaded, enable knowledge base access on any agent:

curl -X PUT "https://api.parallellabs.app/api/v0/agents/{agentId}" \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "knowledgeBaseEnabled": true,
    "knowledgeBaseScope": "/shared/Support"
  }'

Best Practices

Use descriptive file names

Clear file names help with organization and make it easier to manage your knowledge base as it grows.

Keep documents focused

Shorter, topic-focused documents tend to yield more precise retrieval results than large, multi-topic files.

Use folder scoping

Segment documents into folders and use knowledgeBaseScope on agents to keep each agent focused on its relevant content.

Update regularly

Delete outdated documents and upload updated versions to ensure agents always answer with current information.

Get Started

Core Features

Integrations

Platform

Build Your Knowledge Base with Documents

What Are Documents?

How Agents Use Documents

Core Operations

Uploading a Document

Retrieving a Specific Document

Deleting a Document

Enabling Knowledge Base on an Agent

Best Practices

Use descriptive file names

Keep documents focused

Use folder scoping

Update regularly

Next Steps

Documents API Reference

AI Agents

​What Are Documents?

​How Agents Use Documents

​Core Operations

​Uploading a Document

​Listing Documents with Pagination

​Retrieving a Specific Document

​Deleting a Document

​Enabling Knowledge Base on an Agent

​Best Practices

Use descriptive file names

Keep documents focused

Use folder scoping

Update regularly

​Next Steps

Documents API Reference

AI Agents

What Are Documents?

How Agents Use Documents

Core Operations

Uploading a Document

Listing Documents with Pagination

Retrieving a Specific Document

Deleting a Document

Enabling Knowledge Base on an Agent

Best Practices

Next Steps