Generate an answer with text documents

This public endpoint handles chat interactions that involve responses generated from document analysis. It allows users to submit query, which the system processes to provide relevant responses by leveraging documents. You can customize the prompt of the model from the admin panel or complete the existing prompt with a dedicated user message. The query parameter should be a string. The model parameter is optional and specifies the LLM to use for generating responses. If not provided, a default model configured in the user’s chat settings will be used. The workspace_ids parameter is optional and specifies the workspace IDs whose documentswill be added to the scope. The file_ids parameter is optional and specifies the document IDs that will be added to the scope. The chat_session_id parameter is optional and specifies the chat session ID for follow-up. The company_scope parameter is optional and specifies if the documents from the user’s company will beadded to the scope. The private_scope parameter is optional and specifies if the documents from the user’s private collection will be added to the scope. The tool parameter is optional and specifies the tool to use for the chat with doc session choose between ‘VisionDocumentSearch’ and ‘DocumentSearch’, the default is ‘DocumentSearch’. The private parameter is optional and specifies whether this chat interaction is private. The generation_instructions parameter is optional and specifies additional instructions to concatenate with the prompt for generation. These instructions will be appended to the regular prompt. The extra_body parameter is optional and specifies additional parameters for VLLM structured output. Only ‘guided_*’ parameters are allowed (guided_choice, guided_regex, guided_json, guided_grammar, guided_whitespace_pattern, guided_decoding_backend).

POST /api/v2/chat/document-search

Try it

cURL

BasicSearchWithTool

curl --request POST \
  --url https://paradigm.lighton.ai/api/v2/chat/document-search \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "query": "How to implement authentication?",
  "tool": "DocumentSearch"
}
'

200

{
  "chat_session_id": 123,
  "answer": "<string>",
  "user_instructions": "<string>",
  "chat_settings": {
    "company": "<string>",
    "instruction": "<string>"
  },
  "model": "<string>",
  "documents": [
    {
      "id": 123,
      "name": "<string>",
      "document_extracts": [
        {
          "id": 123,
          "metadata": {
            "ID": "3c90c3cc-0d44-4b50-8888-8dd25736052a",
            "pages": "<string>",
            "total_pages": "<string>"
          },
          "text": "<string>"
        }
      ]
    }
  ]
}

Authorizations

Authorization · string · header · required

Bearer token authentication

Body

Request serializer for document search endpoint.

query · string · required

Search query to find relevant documents

tool · enum<string> · default: DocumentSearch

Available options: DocumentSearch, VisionDocumentSearch

private · boolean · default: false

Private search flag

company_scope · boolean · default: false

Include documents from company scope

private_scope · boolean · default: false

Include documents from private scope

workspace_ids · integer[]

List of workspace IDs whose documents will be added to scope

file_ids · integer[]

List of document IDs that will be added to scope

model · string · default: alfred-4.2

Model to use for document search

chat_session_id · integer | null

ID of the chat session for context

Response

200

application/json

Search results

Response serializer for document search results.

chat_session_id · integer · required

answer · string · required

user_instructions · string · required

chat_settings · object · required

Response serializer for light chat settings.

Show child attributes

chat_settings. company · string · required

chat_settings. instruction · string · required

model · string · required

documents · object[] · required

Show child attributes

documents. id · integer · required

documents. name · string · required

documents. document_extracts · object[] · required

Show child attributes

documents.document_extracts. id · integer · required

documents.document_extracts. metadata · object · required

Response serializer for document metadata.

Show child attributes

documents.document_extracts.metadata. ID · string<uuid>

documents.document_extracts.metadata. pages · string

documents.document_extracts.metadata. total_pages · string

documents.document_extracts. text · string