Is it possible to implement a GPT chat in Bubble that will respond based on my database in Bubble?
it really depends on your appâs use case. In a very limited fashion you could simple grab whatever information from the DB is relevant to the user and send it as part of your API request to OpenAI.
For bigger applications youâll want to look into RAG frameworks and MCP servers
Yes it is possible and it exists.
Do these cache the data, or does each query trigger a new database fetch?
No cache per se, but conversation context is leveraged.
For instance, if you ask the details of a testimonial which return the email address of the user, then ask the email address, it will recognise data is already available and will likely answer without a new fetch, although it might query just in case data has changed in between.
This topic was automatically closed after 70 days. New replies are no longer allowed.