Bineric API Documentation

The Deep Research API provides AI-powered comprehensive web research capabilities. It performs extensive web searches on any topic, analyzes the information, and returns a detailed report with proper citations. This feature is fully integrated into the Unified API with both standard and streaming response options.

The Deep Research API supports two response types:

1. Standard Response (Non-Streaming)

   When using the default stream: false option, the API returns a single JSON response after the research completes:

   json

   {
     "success": true,
     "request_id": "req_1234567890_abcdef",
     "model": "Deep Research",
     "response": "Detailed research analysis...",
     "citations": [
       {
         "title": "Source Title",
         "url": "https://example.com/article",
         "snippet": "Relevant excerpt from the source"
       }
       // More sources...
     ],
     "activities": [
       {
         "type": "searching",
         "message": "Searching for latest quantum computing research"
       }
       // More activities...
     ],
     "deep_research_info": {
       "usage_count": 5,
       "usage_limit": 50,
       "usage_left": 45
     },
     "usage": {
       "prompt_tokens": 20,
       "completion_tokens": 1500,
       "total_tokens": 1520,
       "total_cost": 0.25,
       "response_time_ms": 12345,
       "account_info": {
         "type": "paid",
         "balance": 9.75,
         "model_cost": {
           "cost_per_call": 0.25,
           "pricing_model": "fixed_cost_per_call"
         }
       },
       "current_usage": {
         "requests_made": 6,
         "total_input_tokens": 120,
         "total_output_tokens": 9000,
         "total_cost": 1.5
       }
     }
   }

2. Streaming Response

   When using stream: true, the API returns a stream of Server-Sent Events (SSE) with the following event types:

   a. Info Event (start of research)

   javascript

   event: info
   data: {
     "request_id": "req_1234567890_abcdef",
     "model": "Deep Research",
     "message": "Deep research started"
   }

   b. Activity Events (progress updates)

   javascript

   event: activity
   data: {
     "type": "searching",
     "message": "Searching for latest quantum computing research",
     "request_id": "req_1234567890_abcdef"
   }

   Activity types include:

   • search: Information about search queries and results
   • analyze: Updates about analysis of search results
   • synthesis: Status of final report generation

   c. Complete Event (final results)

   javascript

   event: complete
   data: {
     "success": true,
     "request_id": "req_1234567890_abcdef",
     "model": "Deep Research",
     "response": "Detailed research analysis...",
     "citations": [
       {
         "title": "Source Title",
         "url": "https://example.com/article",
         "snippet": "Relevant excerpt from the source"
       }
     ],
     "activities": [...],
     "deep_research_info": {
       "usage_count": 5,
       "usage_limit": 50,
       "usage_left": 45
     },
     "usage": {
       "prompt_tokens": 20,
       "completion_tokens": 1500,
       "total_tokens": 1520,
       "total_cost": 0.25,
       "response_time_ms": 12345,
       "account_info": {
         "type": "paid",
         "balance": 9.75,
         "model_cost": {
           "cost_per_call": 0.25,
           "pricing_model": "fixed_cost_per_call"
         }
       },
       "current_usage": {
         "requests_made": 6,
         "total_input_tokens": 120,
         "total_output_tokens": 9000,
         "total_cost": 1.50
       }
     }
   }

   d. Error Event (if an error occurs)

   javascript

   event: error
   data: {
     "success": false,
     "request_id": "req_1234567890_abcdef",
     "error": {
       "type": "ProcessingError",
       "message": "An error occurred during research",
       "code": 500
     }
   }

The API returns structured error responses with the following format:

{
  "success": false,
  "request_id": "req_1234567890_abcdef",
  "error": {
    "type": "ErrorType",
    "message": "Error message details",
    "code": 400
  }
}

• Fixed price: $0.25 USD per call
• Usage limits:
  • Regular users: 50 calls per month
  • Team accounts: 70 calls per month
• Unused calls do not roll over to the next month

1. Use Streaming for Long Queries: For complex research topics that may take more than a few seconds, use streaming mode to provide real-time feedback to users
2. Handle Errors Gracefully: Implement proper error handling to provide helpful feedback to users when research fails
3. Limit Query Length: Keep research queries clear and focused for best results
4. Manage Rate Limits: Track the user's remaining Deep Research usage and inform them when they're running low
5. Display Citations Properly: Always display source citations alongside research results for proper attribution

• Maximum of 5 parallel search queries by default
• Research may take up to 3 minutes for complex topics
• Usage is limited to 50/70 calls per month depending on account type
• Research is conducted only in languages supported by the search engine