Questions and answers
Q&A pairs: the question is embedded, the answer is returned — the right format for FAQ.
Updated on Aug 10, 2026
For FAQ, embedding the question yields much better search results than embedding running text: the user's question matches directly with the registered question. The Q&A pair does exactly that — the question becomes the vector, and the search returns the answer in the text field, with the original question in question for reference.
https://api.hinow.ai/v1/rag/documents/qaBearerRegister a question-answer pair in the database.
Parâmetros
rag_idstring· bodyobrigatórioquestionstring· bodyobrigatórioanswerstring· bodyobrigatório
Respostas
{
"document_id": "0388772e3fe94ec19217790846d79133",
"chunks": 1,
"embed_tokens": 4
}https://api.hinow.ai/v1/rag/documents/qa/{document_id}BearerEdit an existing pair (re-embeds the question).
Parâmetros
document_idstring· pathobrigatóriorag_idstring· bodyobrigatórioquestionstring· bodyobrigatórioanswerstring· bodyobrigatório
Respostas
In search, a Q&A result arrives like this — notice that text is the answer and the score is high when the question matches:
{
"score": 0.824,
"document_id": "0388772e3fe94ec19217790846d79133",
"source": "Voces aceitam pix?",
"question": "Voces aceitam pix?",
"text": "Sim! Pix, cartao de credito, boleto e transferencia.",
"metadata": { "kind": "qa", ... }
}Removing a pair
Q&A is a document like any other: use DELETE /v1/rag/documents/{document_id}?rag_id=... and it goes to trash like the rest.