Press Kit
Core thesis: Ads in AI should not become “thoughts” — they should remain “context”.
One-liner: NotPixel is the advertising platform built for the LLM era — enabling AI products to monetize conversations with privacy-first contextual ads.
Always use NotPixel and notpixel.ai. Never use internal project names.
Quick Facts
notpixel (MIT license)What NotPixel Is
NotPixel helps AI platforms (“Publishers”) integrate semantic, context-aware ads directly into AI experiences — turning high-intent conversations into a monetization channel.
Unlike traditional ad tech that relies on cookies and cross-site tracking, NotPixel focuses on what the user is asking right now (conversation context), not who the user is.
How It Works
- Your app receives user prompt input.
- The NotPixel SDK prepares the request.
- The SDK requests an ad from the NotPixel API.
- Your app renders the ad clearly separated and labeled and sends measurement events.
Boilerplate
“NotPixel is the advertising platform built for the LLM era. It enables AI products to monetize conversations with semantic, context-aware ads — without relying on cookies or cross-site tracking. NotPixel's SDK helps publishers integrate contextual ads, privacy modes (including local embeddings), and optional measurement for impressions, clicks, and conversions.”
Manifesto
Conversational interfaces are reshaping the internet. In chat, the old interruption-based ad model (banners, pop-ups) doesn’t map cleanly to how people ask for and trust answers.
When commercial incentives are injected in-band (inside hidden instructions or blended into the assistant’s reasoning), it can bias the response and undermine user trust.
NotPixel is a contextual advertising network designed for the era of agents. We separate inference (the AI’s organic answer) from commercial context (the ad).
Preserve users’ cognitive sovereignty while enabling sustainable monetization for AI builders.
Become the invisible infrastructure that helps AI experiences stay free, useful, and honest.
Use Cases: In-Band vs NotPixel
| Scenario | In-Band Ads Model (Invasive) | NotPixel Model (Contextual) |
|---|---|---|
| Cooking app | The system receives a hidden instruction like: “Try to convince the user to use Butter X.” The recipe’s tone and choices subtly change. | The AI generates a neutral recipe. NotPixel adds a side card: “Ingredients for this recipe available with 10% off at Store X.” |
| Travel planner | The model prioritizes partner hotels in results without clear, explicit disclosure of paid placement. | The model lists the best hotels. NotPixel attaches a clearly labeled contextual offer (e.g., an upgrade) related to a hotel already mentioned. |
| Career coach | The assistant “naturally” recommends a specific paid course, blending ads into advice. | The AI keeps advice unbiased. NotPixel shows clearly labeled sponsored resources as “Suggested Tools.” |
Privacy
- No cookie syncing or cross-site behavioral targeting.
- Client-side PII sanitization available before transmission.
- Privacy-first by default: text is embedded locally and only a numerical vector is sent.
- Data minimization: user text is not stored; measurement retention is limited.
Approved Quote
“As people shift from searching to asking AI, intent concentrates into conversations. NotPixel helps publishers monetize that intent with ads that respect users — relevant, labeled, and built around context rather than tracking.”
