Ask LLM

post

https://api.edenai.run/v2/aiproducts/askyoda/v2/{project_id}/ask_llm/

Retrieve a list of search query responses and compare them to your input. Provide a query, and in return, receive scores for the most relevant items from your project, ranked by their proximity to your query.

Recent Requests

Time	Status	User Agent
Retrieving recent requests…

Loading…

Path Params

project_id

uuid

required

Body Params

query

string

required

length ≥ 1

Enter your question or query about the data. The large language model (LLM) will provide a response.

llm_provider

string

length ≥ 1

Select a provider for the large language model for processing. Leave empty for default.

llm_model

string

length ≥ 1

Specify the model to use for language processing. Leave empty for default.

integer

≥ 1

Defaults to 3

How many results chunk you want to return

history

array of objects

Defaults to

A list containing all the previous conversations between the user and the chatbot AI. Each item in the list should be a dictionary with two keys: 'user' and 'assistant'.

history

chatbot_global_action

string

length ≥ 1

A system message that helps set the behavior of the assistant.

filter_documents

object

Filter uploaded documents based on their metadata. Specify key-value pairs where the key represents the metadata field and the value is the desired metadata value. Please ensure that the provided metadata keys are available in your database.

min_score

double

Defaults to 0

A minimum score threshold for the model to consider a chunk as a valid response. Higher values mean the model will be more conservative and only return chunks that are more similar to the query. Lower values mean the model will be more open to returning chunks that are less similar to the query.

temperature

double

0 to 2

Defaults to 0

Higher values mean the model will take more risks and value 0 (argmax sampling) works better for scenarios with a well-defined answer.

max_tokens

integer

≥ 1

Defaults to 1024

The maximum number of tokens to generate in the completion. The token count of your prompt plus max_tokens cannot exceed the model's context length.

conversation_id

uuid

use_reranking

boolean

Defaults to false

Activate/deactivate the reranking. This is experimental.

top_n

integer

Defaults to 3

The number of documents returned by the reranker (experimental)

Responses

200

400

403

404

500