Agents

The Agents API enables interaction with custom AI agents in Egnyte. These agents can be configured with instructions, workflows, and domain-specific knowledge.

This API follows an asynchronous execution model:

1. Submit a question
2. Receive a requestId
3. Poll for the final response

When to Use

Use the Agents API when:

• You need multi-step reasoning or workflows
• You want agent-level customization (instructions, behavior)
• You require conversation continuity across queries

Base URL

https://{domain}.egnyte.com/pubapi/v1/ai/agents

Authentication

All requests require an OAuth 2.0 Bearer token in the Authorization header:

Authorization: Bearer {access_token}

See Authentication for details on obtaining a token.

Rate Limits

The Agents API enforces stricter rate limits than standard Egnyte APIs. These limits apply per access token:

If your production application will exceed these limits, contact Egnyte to discuss custom arrangements.

Execution Flow (Important)

1. List agents

2. Ask question → get requestId

3. Poll status API

4. Retrieve final response

List Agents

Returns all available agents.

Request

GET /pubapi/v1/ai/agents/list

Query Parameters

Example Request

curl -i -X GET "https://{domain}.egnyte.com/pubapi/v1/ai/agents/list?sortBy=name&sortOrder=asc" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

Response

200 OK

Returns an array of agent objects.

Example Response

[
  {
    "agentId": "f52a1253-0b4c-4b1f-a1a6-9dafdd3f3097",
    "name": "llm agent",
    "description": "llm details",
    "status": "ACTIVE",
    "subType": "USER_DEFINED",
    "category": "GENERIC",
    "instruction": "llm details",
    "createdBy": "John Millar"
  },
  {
    "agentId": "bc8223c3-a58d-4fd6-8fc2-072dbc074dc9",
    "name": "test agent 1",
    "description": "sonar issues",
    "status": "ACTIVE",
    "subType": "USER_DEFINED",
    "category": "GENERIC",
    "instruction": "sonar issues only",
    "createdBy": "John Millar"
  }
]

Ask a Question to an Agent

Submits a question to a specific agent and returns a request ID for polling the response. You can optionally provide file context, conversation history, and custom instructions.

Request

POST /pubapi/v1/ai/agents/{agentId}/ask

Path Parameters

Request Body

Example Request

curl -i -X POST "https://{domain}.egnyte.com/pubapi/v1/ai/agents/{agentId}/ask" \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
     -d '{
  "question": "Perform the data analysis and prepare a report"
}'

Example Request with Context

curl -i -X POST "https://{domain}.egnyte.com/pubapi/v1/ai/agents/{agentId}/ask" \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
     -d '{
  "question": "What are the key findings?",
  "conversationId": "b5ec8d41-fc7e-4b20-b913-c84614db6a28",
  "chatHistory": {
    "messages": [
      {
        "role": "user",
        "content": "Hello tell me about llm"
      },
      {
        "role": "assistant",
        "content": "LLMs are..."
      }
    ]
  },
  "entryIds": [
    "{entryId}"
  ],
  "selectedItems": {
    "folders": [],
    "files": [
      {
        "entryId": "{entryId}",
        "filePath": "/Private/asinghi/The Lucy - Transaction Summary (1).docx"
      }
    ]
  }
}'

Response

200 OK

Returns identifiers for polling the agent's response.

Example Response

{
  "requestId": "12c02713-7e79-4a9a-8802-908d93feaa9d",
  "conversationId": "b5ec8d41-fc7e-4b20-b913-c84614db6a28"
}

400 Bad Request

{
  "errorMessage": "Question is mandatory",
  "errorCode": "QUESTION_MANDATORY"
}

500 Internal Server Error

{
  "errorMessage": "An unexpected error occurred. Please try again later.",
  "errorCode": "INTERNAL_SERVER_ERROR"
}

Get Agent Execution Status

Retrieves the status and response of a previously submitted agent question.

Request

GET /pubapi/v1/ai/agents/{agentId}/ask/{requestId}/status

Path Parameters

Example Request

curl -i -X GET "https://{domain}.egnyte.com/pubapi/v1/ai/agents/{agentId}/ask/{requestId}/status" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

Response

200 OK

Example Response

{
  "status": "COMPLETED",
  "responseText": "Large Language Models (LLMs) are a type of artificial intelligence program that can recognize and generate text, among other tasks. They are deep learning models trained on immense amounts of data, enabling them to understand and generate natural language and other types of content to perform a wide range of tasks.",
  "citations": [
    {
      "previewUrl": "https://www.cloudflare.com/learning/ai/what-is-large-language-model/",
      "type": "WEB_SEARCH"
    },
    {
      "previewUrl": "https://en.wikipedia.org/wiki/Large_language_model",
      "type": "WEB_SEARCH"
    }
  ],
  "lastUpdated": "2025-09-18T10:24:59Z"
}

Error Codes

Code Examples

POST /pubapi/v1/ai/agents/{agentId}/ask

curl -i -X POST "https://{domain}.egnyte.com/pubapi/v1/ai/agents/{agentId}/ask" \
     -H "Content-Type: application/json" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
     -d '{
  "question": "Perform the data analysis and prepare a report"
}'

GET /pubapi/v1/ai/agents/{agentId}/ask/{requestId}/status

curl -i -X GET "https://{domain}.egnyte.com/pubapi/v1/ai/agents/{agentId}/ask/{requestId}/status" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

GET /pubapi/v1/ai/agents/list?sortBy=createdOn&sortOrder=desc

curl -i -X GET "https://{domain}.egnyte.com/pubapi/v1/ai/agents/list?sortBy=createdOn&sortOrder=desc" \
     -H "Authorization: Bearer YOUR_ACCESS_TOKEN"

Related Resources

• AI API — General AI capabilities for content analysis and generation
• Authentication — How to obtain and refresh OAuth tokens
• Best Practices — Rate limiting, error handling, and optimization strategies
• File System API — Access files and folders to use as agent context