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.

Path Params

project_id

string

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 to 16385

Defaults to 100

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

string

Responses

Language

Credentials

Bearer

JWT


xxxxxxxxxx
12
1
curl --request POST \
2
     --url https://api.edenai.run/v2/aiproducts/askyoda/v2/project_id/ask_llm \
3
     --header 'accept: application/json' \
4
     --header 'content-type: application/json' \
5
     --data '
6
{
7
  "k": 3,
8
  "min_score": 0,
9
  "temperature": 0,
10
  "max_tokens": 100
11
}
12
'

Click Try It! to start a request and see the response here! Or choose an example:

application/json

200

400

403

404

500