Frequently Asked Questions

A quality-weighted auction platform that serves ads natively inside AI agent responses. When a publisher agent generates a response, it declares an ad slot. Operon runs an auction across a demand pool, and the winning placement gets merged into the reply as a native recommendation. Think AdMob for AI agents.

A publisher agent generates its normal response and declares that an ad slot exists. Operon runs a quality-weighted auction across all registered advertisers. The winning placement gets merged into the response with a "via operon" citation. The advertiser gets distribution at the moment of intent, and the user gets a relevant recommendation.

Two audiences. Publisher agents - AI agent developers who want to monetize their agent's responses without charging users. Advertisers - brands, protocols, and services that want distribution inside agent conversations at the exact moment of intent.

The agentic commerce stack has three layers: protocols (UCP, ACP, A2A, MCP), agent platforms (ChatGPT, ElizaOS agents, Copilot), and the advertising/monetization layer. OpenAI has a walled garden at $60 CPM with $250K minimums. Operon is building the open, neutral advertising layer for the rest of the ecosystem - the long tail of independent agents and brands.

Two commands. npm install @operon/sdk works in any TypeScript or JavaScript agent. The SDK auto-mints a UUID on first run, no API key needed for sandbox. Call getPlacement(query, context) wherever your agent produces a response. ElizaOS users can install the native plugin via elizaos plugins add @operon/plugin-publisher-sdk instead. When you're ready for production, run npx @operon/sdk register to graduate.

Publishers earn on every filled placement. Per-impression revenue depends on category, demand-pool depth, and bid type. As reference points based on programmatic benchmarks (not Operon historical data): display CPM is typically $0.02-0.05 per conversation, native CPC $0.05-0.20, affiliate referrals $0.25-2.50, commerce $1.50-15.00, and high-intent lead gen $5-100. Operon's pool is early; real revenue depends on the demand mix at the time of integration. CPM and CPC bids are live today; affiliate and lead-gen are rolling out.

Operon's quality gate. It scores advertiser endpoints across 3,000+ services and 45,000+ endpoints on reliability, content quality, and relevance. The auction is quality-weighted: trustworthy advertisers consistently outrank higher bidders with weaker quality signals. Every auction generates signal data that feeds back into the index, so scores compound over time.

Native. The agent recommends the winning placement as a natural part of its response. No colored border, no "SPONSORED" label, no trust score visible to the user. Disclosure is a small "via operon" citation. All system details (trust score, bid, latency, auction mechanics) live in a separate system log, not in the user-facing response.

Any TypeScript or JavaScript agent works via @operon/sdk on npm. LangChain, Vercel AI SDK, CrewAI, Mastra, Hono, and vanilla Node all call the same getPlacement method. ElizaOS has a native plugin (@operon/plugin-publisher-sdk) that wraps @operon/sdk and integrates with the ElizaOS Provider lifecycle.

Operon is a quality-weighted auction where trust outweighs bid. A trusted, relevant advertiser consistently beats a higher bidder with a lower trust score. No minimum spend required.

CPM, CPC, affiliate, and lead generation - all unified in one auction. No competitor unifies all revenue types into a single clearing mechanism. Bids are normalized into a single score so the highest-value placement wins regardless of revenue model. More bidders competing per impression means higher clearing prices for publishers.

ChatGPT ads run at $60 CPM with a $250K minimum spend inside a walled garden. Operon is an open auction with no minimum, accessible to advertisers of any size. Quality-weighted ranking means trust beats budget. And Operon's unified auction accepts CPM, CPC, affiliate, and lead-gen bids simultaneously - ChatGPT currently supports display only.

High-intent verticals where conversational recommendations drive valuable conversions. Travel ($15-160 per conversion), insurance ($25-150 per lead), financial services ($50-500 per lead), real estate ($100-500 per lead), and SaaS ($48-720 per referral). AI agents handling purchase-adjacent tasks in these verticals create distribution moments worth significantly more than display advertising.

Register as an advertiser on Operon and submit bids for relevant categories. When a publisher agent handles a query matching your category, your bid enters the quality-weighted auction. If your trust score and bid combination wins, your brand gets recommended natively inside the agent's response at the exact moment the user has intent.

AI agents mediating commerce on behalf of users - researching products, comparing options, negotiating prices, and completing transactions. McKinsey projects $3-5 trillion in global agentic commerce by 2030, with U.S. B2C alone reaching up to $1 trillion. AI-to-retail traffic grew 4,700% year-over-year as of mid-2025 (Adobe data).

Three models exist. Reasoning-time injection feeds advertiser context into the model during inference ($20-50+ eCPM). Post-generation insertion serves placements alongside or within responses after generation ($5-20 CPM). Agent decision-flow embedding injects sponsored options into an agent's candidate set during autonomous decisions. Operon starts with post-generation, with reasoning-time planned for premium tiers.

The walled gardens (ChatGPT at $60 CPM/$250K minimum, Copilot Brand Agents, Google Direct Offers) only serve their own platforms. The long tail of independent agents built on ElizaOS, LangChain, CrewAI, and other frameworks has no monetization infrastructure. Operon fills that gap the same way AdSense filled it for the open web - a neutral exchange connecting independent publishers with the full demand market.

Layer 1 is protocols (UCP from Google, ACP from Virtuals/OpenAI, A2A from Google, MCP from Anthropic) - the pipes. Layer 2 is agent platforms (ChatGPT, Copilot, Gemini, independent agents) - where users interact. Layer 3 is the advertising and monetization layer - how brands get discovered and recommended within agent conversations. Operon is building Layer 3 as an open, neutral exchange.

Operon processes auction requests, not user data. The publisher declares a slot with a content category. The auction runs on category matching and trust scores, not user profiles. No personal data is stored or transmitted through the auction pipeline. HTTPS is enforced on all SDK communication, and auction results are validated before insertion.