An AI agent that answers from what you actually know, not what it guesses
Most WhatsApp “AI chatbots” either parrot a generic script or hallucinate an answer. Ours drafts replies grounded in your FAQs, policies, and product docs — using your own OpenAI or Anthropic key — and hands off to your team the moment it's unsure.
Guessing ≠ Grounded
Why chatbots that make things up frustrate customers, and honest hand-offs build trust instead.
"Delivery usually takes a few days, don't worry!" (made up on the spot, not from your real policy)
"Yes, you're all set for 4pm!" (no calendar was ever checked)
"I'm sorry, please try resetting the device." (wrong category of problem entirely)
"Standard delivery is 3–5 business days after dispatch. I don't have live courier tracking — if you need the exact status I'll connect you with our team."
"I can't access the live appointment calendar myself, so I'm connecting you with our front desk to confirm Dr. Sharma's 4pm slot."
"I'm really sorry to hear that — connecting you with our support team right now with your message attached."
See how it drafts from your knowledge base
Click a scenario below to see how a customer message becomes a grounded draft, an auto-reply, or a clean hand-off.
VyaparPing AI Agent Playground
Knowledge-Base GroundedTrained on your documents. It drafts grounded replies and hands off the moment it's unsure — it never guesses.
Hi, can I return a kurta I bought 5 days ago? It doesn’t fit right.
“Items can be returned within 7 days of delivery if unused and with original tags attached.”
Yes! Since it’s been 5 days, you’re within our 7-day return window. Please share your order ID and I’ll start the return for you.
You bring the key. You set the guardrails.
Draft-only or auto-reply, a reply cap per conversation, and exactly where it hands off when it's unsure — all in your control.
Bring Your Own Key
Connect your own OpenAI or Anthropic API key. VyaparPing calls the provider directly — no per-seat AI fee, and your key and data stay yours.
Trained on Your Data
Upload PDF catalogs, pricing sheets, return policies, and FAQ docs. The assistant retrieves the relevant passage before it drafts a reply.
Hindi, English & Hinglish
Understands Indian conversational dialects, colloquialisms, and transliterated roman script out of the box.
What it won't do
It will never invent an answer. If it cannot ground a reply in your verified knowledge base, or it hits the reply cap you set, it says so plainly and hands over to a human agent with the entire chat history attached. We would rather it hand over to a human than confidently tell your customer something false.
Further reading
Background guides on the setup, pricing and mechanics behind this page.