An AI support agent grounded in your own content that answers visitor questions with cited sources, captures a lead when it can't help, and escalates to a human when it's unsure. It deflects the repetitive tickets so your team handles the ones that matter.
What the support agent does
Most support questions are already answered somewhere in your docs, help center, or product pages. The support agent indexes that content and answers instantly — with a citation to the exact page, so visitors trust the reply. It runs 24/7, in the visitor's language, and never invents an answer: when the information isn't in your content, it says so and offers to take an email or escalate.
Grounded, not guessing
The agent only answers from content you indexed — a URL crawl, a GitHub repo, uploaded PDFs, or a connected MCP source. That grounding is the difference between a helpful assistant and a liability. Every answer links to its source, and the analytics dashboard surfaces the questions it couldn't answer, so you know exactly which docs to write next.
Actions, with approval
Through the Model Context Protocol (MCP), the support agent can do more than talk — create a ticket, look up an order, or update a record — but anything that changes data pauses for human write-confirm first. It also hands off to a Sales or Technical agent when the conversation shifts, part of the broader multi-agent platform.
Getting started
Index your content, enable the support agent, and embed one script tag. Try it live in the playground, then start free — 100 messages a month on the free tier, with room to grow on pricing.
Will it make up answers?+
No. The support agent answers only from content you indexed and cites the source. When the answer isn't there, it says so and offers to capture a lead or escalate.
Can it do more than answer?+
Yes. Via MCP it can take actions like creating a ticket or looking up an order — but any write action requires human approval first.
Is there a free plan?+
Yes — 100 messages a month and one website on the free tier. See pricing for higher limits.