Configure persona and tools
Choose a 3D avatar, write your agent's system prompt, upload knowledge documents, add paid tools from the marketplace, and set a voice. All in the visual editor.
Configure your agent's 3D avatar, voice, knowledge base, and paid tools in the Widget Studio — then copy a single <script> tag that runs on any website, CMS, or app. No back-end required.
How it works
Choose a 3D avatar, write your agent's system prompt, upload knowledge documents, add paid tools from the marketplace, and set a voice. All in the visual editor.
Chat with your agent live in the preview pane. Adjust the avatar, tune the prompt, change the voice — the preview updates without refreshing. Test edge cases before shipping.
When you're ready, copy the generated <script> tag and paste it anywhere — Webflow, WordPress, a custom HTML page, or a React app. The widget appears as a floating button.
The floating widget opens a full 3D avatar that speaks and lip-syncs — not a flat chatbot bubble.
Choose from dozens of voices or clone a custom voice in Voice Lab and assign it to the widget. Text responses are read aloud with lip-sync.
Upload PDFs, paste URLs, or connect a sitemap. The agent retrieves from your documents before answering — accurate, on-topic replies out of the box.
Attach x402 paid skills — market data, code execution, image generation — and the agent charges callers per-use automatically via Solana micropayments.
Configure in minutes. No code, no back-end. One script tag.