Website chatbot (RAG)

Answers visitors' questions from your own site and docs — accurately, with sources — and captures leads while it's at it.

What it does

• Answers from your website, docs, and knowledge base
• Cites where answers come from
• Captures leads and contact details
• Hands off to a person when needed

Common requests it handles

• What does [feature] do?
• How is your product different from X?
• Do you support [use case]?

Recommended models

RAG quality depends more on retrieval than raw model size — a mid-size open model (Llama, Qwen, Gemma) grounded in your docs answers accurately; an API model helps for longer, synthesis-heavy answers.

Tuning tips

• Keep your source docs clean and up to date — garbage in, garbage out
• Require citations so answers stay grounded
• Capture leads at the right moment, not on every message

What we need from you

• Your site, docs, or knowledge base
• What counts as a lead
• Where leads should go

Good for

• Content-rich sites and SaaS
• Deflecting repetitive questions
• Turning traffic into leads