Peak intent: a developer waiting on their tooling
Most developer advertising fights for scraps of attention — a sidebar on a docs page, a slot in a newsfeed someone is half-scrolling. The wait-state is different. When an AI coding agent is thinking, the developer has stopped, is staring at the screen, and has a few idle seconds they did not plan for.
That is exactly the audience you want to reach: software developers, in their editor, mid-flow, deciding what tools to pull into the stack they are building right now. If you are figuring out how to advertise to software developers without being the ad they instantly block, this is the moment that actually converts.
The format: one clean line, ~5 seconds, non-blocking
Your ad is a single line of text that appears next to the agent's spinner for about five seconds per impression. It never blocks the developer's work, it never interrupts the agent, and the developer opted in — they can flip it off with one click. No takeover, no autoplay, no dark patterns.
Because it is one line, the bar for the creative is high and the tolerance for noise is low. That constraint is the point. Developers will read a line that respects their time. They will not forgive a line that wastes it, so moderation is strict (more on that below).
A live bid market, not a fixed-CPM rate card
There is no sales call and no negotiated annual minimum. You bid in a live market, in blocks of 1,000 impressions, and the highest bid takes the #1 spinner slot. Spend follows demand: when you want the top line, you bid for it; when you don't, you stop.
As a reference point, the top slot tends to clear around a $20 CPM. About half of that reaches the developer rendering your line — which is the deal that keeps the supply side honest and the inventory real.
- ▸Bid per 1,000 impressions; the highest bid wins the #1 slot.
- ▸No invite-only gatekeeping and no five-figure monthly commitment to get started.
- ▸Set your spend and pacing; pause anytime. Settlement runs in USDC on Solana.
Every impression is a real developer-second (no bot farms)
An impression here is a person watching a spinner, not a pixel loaded in a headless tab. The line renders inside the developer's editor while a real AI coding session is genuinely waiting, and the extension reports that a view happened. That is the whole event.
The privacy model is part of why the inventory is clean: the extension renders the ad and reports that a view occurred — it never reads code, prompts, or files. So you reach developers in their IDE without ever touching what they are working on, and without buying traffic that no human ever saw.
Moderated creative, controlled spend and pacing
One bad line poisons the channel for everyone, so creative is moderated before it runs. Keep it factual and useful — what your product does, who it is for — and skip the hype. Lines that read like real recommendations from one engineer to another perform; lines that read like a billboard get rejected.
On the buy side you stay in control: set your bid, cap your spend, and pace impressions so a campaign does not burn out in an hour. Because everything settles on-chain in USDC, your spend and the developer payouts it funds are verifiable on Solscan — this runs on Solana mainnet, with real money on both sides.
How it compares to other developer ad channels
The honest pitch for developer marketing in 2026 is that attention is scarce and ad blockers are universal, so the channels that work are the ones developers tolerate. Privacy-first display networks like EthicalAds and curated programs like Carbon Ads put clean ads on docs sites and dev blogs; daily.dev places ads in the developer news feed. Those reach developers between tasks.
claudearn reaches them during a task, in the editor, at the exact second they are idle and watching. It is a different surface, not a replacement — many advertisers run both. Direct wait-state networks (Kickbacks, CodeCash, WaitJI, Idlen and others) target the same spinner moment; their splits, payouts and privacy models vary, so check each one's current terms before you compare.
- ▸Surface: the AI coding agent's wait-state, inside the editor — not a docs sidebar or a feed.
- ▸Pricing: live per-1,000-impression bidding, no five-figure minimum to start.
- ▸Quality: real developer-seconds, with payouts settled on-chain and verifiable.
Frequently asked questions
Who exactly sees my ad?
Software developers using AI coding agents in VS Code, Claude Code, and Codex, at the moment their agent is thinking and they are waiting on a result. Your line shows for about five seconds per impression next to the spinner. It is opt-in on the developer's side and never blocks their work.
How does pricing work — is there a minimum?
You bid in a live market in blocks of 1,000 impressions, and the highest bid takes the #1 spinner slot. The top slot tends to clear around a $20 CPM. There is no invite-only gate and no large monthly commitment to start; you set your spend and pacing and can pause anytime.
Will you give me data on what developers are coding?
No, and that is by design. The extension renders your ad and reports that a view happened — it never reads code, prompts, or files. You reach developers in their IDE without us collecting or selling what they are working on.
What can the ad actually say?
One clean line of text, moderated before it runs. Keep it factual — what your product does and who it is for. Hype, clickbait, and anything misleading gets rejected, because a single bad line degrades the channel for every advertiser and every developer on it.
How is spend handled and is it verifiable?
Settlement runs in USDC on Solana mainnet, with real money. About half of your CPM reaches the developer rendering your line, and those payouts settle on-chain — verifiable on Solscan. You control your bid, your cap, and your pacing throughout.