API (Agent)
The API (Agent) allows you to interact programmatically with your deployed agent by sending queries and receiving responses. This information is also directly available within the platform.
Query API
Purpose
Send a question to your agent and receive a structured response after it passes through the full agent processing flow.
Endpoint
nginxCopyEditPOST https://atlantis.greatwave.ai/api/query
Parameters
key
Unique public key assigned to the agent
query
User question or message
memory_id
Optional ID to maintain continuity in a session
query_id
Optional session key for current agent conversation
Sample Request
bashCopyEditcurl -X POST https://atlantis.greatwave.ai/api/query \
-H "Content-Type: application/json" \
-d '{
"key": "your-public-agent-key",
"query": "start the evaluation on the latest reports",
"memory_id": "UNIQUE-SESSION-ID"
}'
Sample Response
jsonCopyEdit{
"ref": 1234,
"response": "This is a response from the agent.",
"chunks": [{
"chunk": "Filename.pdf",
"chunk_type": "pdf file used for the response",
"comment": "Metadata attached",
"score": 0.87
}]
}
Query Feedback API
Purpose
Submit reviews, comments, or ratings on the responses returned by your agent. Useful for training insights or performance tracking.
Endpoint
nginxCopyEditPOST https://atlantis.greatwave.ai/api/query-feedback
Parameters
key
Unique public key assigned to the agent
query
Original user input
response
Response generated by the agent
feedback
Freeform text comment (e.g., review or evaluation)
feedback_type
"positive" or "negative"
Sample Request
bashCopyEditcurl -X POST https://atlantis.greatwave.ai/api/query-feedback \
-H "Content-Type: application/json" \
-d '{
"key": "your-agent-key",
"query": "start the evaluation",
"response": "This response was really good.",
"feedback": "This response was really good.",
"feedback_type": "positive"
}'
Sample Response
jsonCopyEdit{
"message": "Feedback was submitted successfully"
}
This API set allows you to:
Automate agent interaction via backend systems.
Log and analyze performance of responses.
Collect user satisfaction metrics for continuous improvement.
Last updated