How it works
Your bot stores endpoint settings and calls the API at answer-time. The final answer is generated by the LLM using both trained content and live API context.
Live API lets your bot fetch fresh data when a question is asked, then combine it with trained documents before generating the final response.
Not looking for human agent handoff? See Live Chat Handoff to let visitors ask for human help when no AI answer is sufficient.
Your bot stores endpoint settings and calls the API at answer-time. The final answer is generated by the LLM using both trained content and live API context.
Live API is available on Business and Scale plans.
Return filtered and paginated records that match the user query instead of full dataset dumps.
Answer questions about current stock, table openings, service slots, or appointment availability without retraining documents each time data changes.
Return a narrow, permission-safe response from an internal endpoint, then let the bot explain the result alongside your support docs.
Fetch upcoming classes, event times, specials, webinars, or staff calendars while keeping static policies in uploaded sources.
Use an endpoint to calculate a quote, package fit, or eligibility result, then ground the explanation in your existing documentation.
GET /api/availability?service=catering&date=2026-07-12
Authorization: Bearer YOUR_API_TOKEN{
"available": true,
"slots": ["12:00", "14:30", "17:00"],
"minimumGuestCount": 20,
"notes": "Outdoor catering requires 7 days notice."
}Return concise JSON that is already scoped to the visitor's question. Include labels, status values, timestamps, limits, and human-readable notes. Avoid sending secrets, unrelated records, or entire tables; the endpoint should act like a query-specific context provider for the bot.
Explore the chatbot platform
Answer routine support questions and hand off when needed.
Support visitors, leads, and teams with business knowledge.
Turn FAQs and approved knowledge into direct answers.
Turn documents into grounded, searchable answers.
Import site content and answer visitors on-page.
Restrict internal knowledge to approved users.
Combine indexed knowledge with fresh endpoint data.
Answer first, then capture qualified follow-up details.