Question & Answer

Allows to query the knowledge base using a specific AI model.

The Knowledge Base model

Key	Type	Description
id	String	The unique identifier for the knowledge base which is given by Tiledesk.
name	String	The knowledge base name.
id_project	String	The unique identifier of the project
preview_settings	Object	The settings for the knowledge base preview
default	Boolean	Specifies if the knowledge base is the default one
hybrid	Boolean	Specifies if the knowledge base is hybrid. Default is false (standard type)
engine	Object	Specifies the configuration of the vector store system used by the knowledge base. A default engine is provided.
embedding	Object	Indicates which embeddings are used for vector-based search. A default embedding is present.

Ask the Knowledge Base

POST https://api.tiledesk.com/v3/:project_id/kb/qa

Allows to query the knowledge base using a specific AI model.

Path Parameters

Name	Type	Description
project_id	string	The Project Id is a unique code assigned to your project when you create it in Tiledesk.

Headers

Name	Type	Description
Authorization	string	Authorization token. Basic Auth or JWT

Request Body

Name	Type	Description
question	string	The question submitted
namespace	string	The id of the Knowledge Base in which to search for the answer
llm	string	(Optional) The LLM to use to generate the response (Default: "openai")
model	string	The model to use to generate the response (e.g. gpt-4o)
max_tokens	Number	The maximum number of tokens that can be consumed to generate the response
temperature	Number	Defines creativity in generating responses (low values ​​determine more specific and predictable responses)
top_k	Number	The number of nearby chunks to use to generate the response
engine	Object	(Optional) The engine configuration object to use for the vector store system if different from the Default Engine
embedding	Object	(Optional) The embedding object to use for the embedding generation if different from the Default Engine
system_context	string	(Optional) The context to give to the AI ​​to shape its behavior in generating the response.
alpha	Number	(Optional) Defines the balance of hybrid search (0 = full-text focused, 1 = semantic/vector focused).
chunks_only	Boolean	(Optional) Returns only the retrieved text chunks without generating a final answer.
stream	Boolean	(Optional) Enables real-time streaming of the response as it is generated.
citations	Boolean	(Optional) Includes source references for the retrieved or generated content.
tags	Array	(Optional) Filters or categorizes results based on associated tags.
rereanking	Boolean	(Optional) Applies a secondary ranking step to improve the relevance of retrieved results.

200

{
    "answer": "To create an AI assistant using OpenAI, you can follow these steps:\n\n1. **Visit OpenAI**: Navigate to the OpenAI website.\n2. **Access the API Section**:\n   - Go to ‘Products’, then select ‘API’.\n   - Log in and select ‘API’.\n3. **Navigate to Assistant Creation**:\n   - Ensure you are on the Dashboard.\n   - Click on ‘Assistant’ from the left sidebar menu.\n4. **Create the Assistant**:\n   - Click the green ‘Create’ button in the top right corner.\n   - Name your assistant and provide context in the ‘Instructions’ section to fine-tune its responses.\n5. **Select the Model**:\n   - For this example, you can use GPT-4o.\n6. **Handle File Formats**:\n   - If uploading a CSV file, use a Code interpreter.\n   - For PDF or text files, use the File Search feature.\n7. **Integrate with Tiledesk**:\n   - Copy the assistant ID.\n   - Go to the Tiledesk dashboard, click on the block where the ChatGPT Assistant is placed.\n   - In the right-side menu, paste the assistant ID into the “Assign GPT Assistant” field.\n\nIf you need",
    "success": true,
    "namespace": "66a897133eaa7f0013632c5b",
    "id": "66b6268722af86ab6a739cb6",
    "ids": [
        "66b6268722af86ab6a739cb6"
    ],
    "source": "https://gethelp.tiledesk.com/articles/create-an-ai-assistant-in-openai/",
    "sources": [
        "https://gethelp.tiledesk.com/articles/create-an-ai-assistant-in-openai/"
    ],
    "content_chunks": null,
    "prompt_token_size": 1185,
    "error_message": null,
    "chat_history_dict": {
        "0": {
            "question": "how can i create an AI assistant?",
            "answer": "To create an AI assistant using OpenAI, you can follow these steps:\n\n1. **Visit OpenAI**: Navigate to the OpenAI website.\n2. **Access the API Section**:\n   - Go to ‘Products’, then select ‘API’.\n   - Log in and select ‘API’.\n3. **Navigate to Assistant Creation**:\n   - Ensure you are on the Dashboard.\n   - Click on ‘Assistant’ from the left sidebar menu.\n4. **Create the Assistant**:\n   - Click the green ‘Create’ button in the top right corner.\n   - Name your assistant and provide context in the ‘Instructions’ section to fine-tune its responses.\n5. **Select the Model**:\n   - For this example, you can use GPT-4o.\n6. **Handle File Formats**:\n   - If uploading a CSV file, use a Code interpreter.\n   - For PDF or text files, use the File Search feature.\n7. **Integrate with Tiledesk**:\n   - Copy the assistant ID.\n   - Go to the Tiledesk dashboard, click on the block where the ChatGPT Assistant is placed.\n   - In the right-side menu, paste the assistant ID into the “Assign GPT Assistant” field.\n\nIf you need"
        }
    }
}

Example

curl -v -X GET -u [e-mail]:password -d {"question":"how can i create an AI assistant?","namespace":"66a897133eaa7f0013632c5b","model":"gpt-4o","temperature":0.7,"max_tokens":256,"top_k":4,"system_context":null} https://api.tiledesk.com/v3/63ad512e70d5ed0012ad6286/kb/qa

---