MailsRAG

first to comment, great product

great use of automation. really helpful

MailsRAG is built for teams that handle repetitive support and sales email but do not want raw model output going straight to customers. It combines knowledge retrieval, AI drafting, and human review so teams can automate email in a more reliable way. The best fit is support and sales teams with existing docs, FAQs, policy pages, or internal knowledge that can ground replies.

Great idea, the utilisation are really good, i will use it

Nice execution on a real pain point. Review-first by default is the right call for LLM-drafted customer replies, to build trust before going full-auto. RAG-grounded answers + multilingual support should keep quality consistent across agents. 30 seconds vs. 3-5 minutes per email is a strong number if it holds up. Curious how it handles questions outside the KB, is it by human handoff?

Keeping a human review step between the draft and the customer is the part most email automation skips. Does the retrieval pull from live docs or a snapshot?

MailsRAG strikes a great balance between AI automation and human oversight. Grounding responses in trusted company knowledge while keeping a review step makes it much more practical for support and sales teams that need both speed and accuracy.

Users may receive customer emails in various languages, and the system supports all languages, including very small language families, to ensure consistency between the reply language and the email source language. This feature is already technically easy to implement and brings great convenience to users

MailsRAG Your product has strong potential, but I found a few key improvements that could make it even better. I'd love to share my feedback and suggestions—please contact me at [email protected]