Everything your customer line needs.
The capabilities that turn a chat widget into a semantic PBX. Pick a card to read deeper, or jump straight to demo or pricing.
Six pillars, one routing layer
Semantic routing
The Sonar intent engine reads message content, conversation context, and policy to decide whether to answer, ask, run a tool, or hand off.
AI human handoff
Confidence triggers, policy escalations, customer intent — all surface to the right human on the endpoint they already use.
Knowledge base AI
RAG that knows when not to answer. Cite the source, link the original document, and ask for a person when confidence drops.
Bring your own keys
Your Meta app, your WABA, your LINE channel, your AI keys. No reseller markup. No lock-in. Your provider relationships stay yours.
Agent-native endpoints
Reply from WhatsApp, Telegram, LINE, Meta, the PWA, or a signed webhook. One active endpoint per agent. No collisions.
Podium live board
Owners and supervisors get a live view of every conversation, routing decision, and team availability across every channel.
Frequently asked questions
What is the most important feature?
Semantic routing is the foundation. Everything else — handoff, knowledge, endpoints, Podium — is either an input or an output of the routing decision.
Can I disable AI answering and only use semantic routing?
Yes. Configure the routing policy to never answer directly and always ask for a clarifying signal or hand off. The Sonar intent layer still does the work.
Do features differ between channels?
The platform exposes a consistent feature surface across channels, but each channel has its own constraints. WhatsApp templates, Instagram 24-hour windows, and LINE Official Account quotas are honored.
Keep exploring
Ready to give your customers one place to talk to you?
Start free on your own site in minutes. Connect your channels later. Bring your own keys when you do.