No-code quickstart
Use the dashboard to create the workspace, test key, disclosure label, and first chat placement before touching production traffic.
AI chatbots and assistants
Monetize chat moments without changing the answer.
Add clearly labeled sponsored moments around assistant replies, search results, and recommendations while your model response, prompt, and conversation history stay untouched.
The sponsored moment
Monetize the visible AI state.
Chat products create visible states where users are waiting for, reading, or acting on an assistant response. wavebird monetizes that chat surface, not the chatbot answer.
While the model is generating
Directly below an assistant response
Between conversation turns
Beside a response in a companion panel
Near recommendations or search results
Use a clearly labeled slot around the assistant response, especially below long answers or beside search-style responses.
Visual sponsored moments, not rewritten AI output.
Some AI monetization approaches insert sponsored text into model output. wavebird keeps the boundary clear with visual sponsored moments that are labeled, separate, and controlled by your app.
Patent-pending generation-time sponsorship.
wavebird introduced a system for showing sponsored moments while an AI response is being generated, turning wait time into revenue without delaying or altering the answer.
Recommended setup
Start with the smallest controlled surface.
The first test should be easy to reason about: one visual slot, one timing rule, clear disclosure, and no sensitive data in the sponsor path.
Why this surface works
- Chat users often wait for generated responses, creating a natural attention window.
- The answer area already has focus, but the sponsor can sit around it without becoming part of the conversation.
- Free chat users can create high inference cost without subscribing, so controlled monetization matters.
What stays untouched
- Model response
- Prompt
- Conversation history
- System instructions
- Tool calls
- User identity and sensitive details
Placement judgment
Good placements versus avoid.
The difference is whether the sponsor is clearly outside the AI output and easy for the user to understand.
Good
- A clearly labeled sponsored card below a completed assistant answer.
- A companion banner beside a long AI answer.
- A sponsored recommendation module after a search-style response.
Avoid
- Sponsored text inserted into the assistant answer.
- Sponsor instructions added to the prompt or system message.
- Ads that look like part of the model's reasoning or recommendation.
Implementation path
Server-side match, client-side render.
The product keeps orchestration and policy control. wavebird returns a placement decision, then the hosted renderer handles media and beacons.
Developer API path
Call /v1/placements from your backend, load /v1/render.js in the frontend, and wrap the chat turn with wavebird.withTurn().
Server-side matching
Send session id, job type, slot hint, broad category, language, and consent state. Keep the raw message out by default.
Client-side rendering
Render the returned placement in a separate slot below, beside, or between chat turns. Hide the slot on no-fill.
Controls
Decide where sponsorship is allowed.
The app decides surfaces, formats, labels, relevance, blocked categories, fallback behavior, and when the sponsor path is disabled.
Allowed formats
Generation or post-response timing
Eligible chat surfaces
Disclosure label
Blocked industries
Relevance rules
Fallback behavior
Kill switch for sensitive chats
Related surfaces
Similar visible AI moments.
Next step
Build your first sponsored chat moment.
Start with one clearly labeled slot around assistant replies, then expand only where the experience stays clean.