Knowledge base
A chatbot that answers with what's in your help center
Retrieve the right passages from your documentation for every question and let the language model answer with sources, without making things up.
The problem
On their own, language models don't know your product and make up answers. Building your own semantic search takes a vector database, embedding models and ongoing maintenance.
The solution
Upload your articles to a knowledge base and, for every question, retrieve the most relevant passages. The model answers using only that context, and customer feedback improves future searches.
How it works
AI support chatbot
Index your help center
Upload each article in Markdown with its own identifier; updates replace the previous version.
Search on every question
POST /v1/tools/rag/collections/{id}/query with the customer's question.
Answer with sources
Pass the passages to your language model and show a link to the article used.
- Answers grounded in your articles, with sources
- No vector database or models to maintain
- Customer feedback improves search
curl -X POST 'https://api.respawntech.dev/v1/tools/rag/collections/3c9e7a52-1b4d-4f0e-9a6c-2d8e5f7a1b3c/query' \
-H "X-API-Token: $RESPAWN_API_TOKEN" \
-H 'Content-Type: application/json' \
-d '{ "query": "como mudo minha senha?", "top_k": 3 }'FAQ
Frequently asked questions
Which language model should I use?
Whichever you prefer. The API returns the passages; generating the answer is up to the model of your choice.
How do I keep the knowledge base up to date?
Resend the article with the same identifier whenever it changes. The old version keeps answering until the new one is ready.
Start integrating today
Create your account, generate a token and make your first call in a few minutes.