Why in-editor beats banner-blinded channels
Developers are the hardest audience to reach with display advertising, and they earned that reputation honestly. They run ad blockers by default, they ignore docs-site banners, and they treat most marketing as noise to be filtered. The conventional dev-ad channels — display units on documentation sites, curated sidebar placements, ads in a news feed — work around the edges of attention, never at the center of it.
In-editor developer ads invert that. Instead of competing with the content a developer came for, your line appears in the dead seconds between intent and result, while the agent is still thinking. There's nothing else on screen to compete with and nothing being interrupted. It's one line, it never blocks the work, and it's gone in about five seconds.
That's the difference between buying impressions a developer trained themselves to ignore and buying impressions in the one surface they're actually watching.
Where your line appears: the AI coding agent spinner
When a developer runs an AI coding agent, the agent spends real time thinking before it answers. During that wait, the editor shows a spinner. claudearn slips one tasteful sponsored line into that spinner's wait-state — alongside the status, not on top of the work.
It renders across VS Code, Claude Code, and Codex, the terminal CLI. The developer opted in, can reverse it with one click, and is being paid for the view. So the placement is consensual on both sides of the screen: you're not ambushing anyone, you're sponsoring a moment they already agreed to monetize.
- ▸One line, about five seconds, never blocking the developer's task.
- ▸Appears in the AI agent's "thinking" spinner inside VS Code, Claude Code, and Codex.
- ▸Opt-in and one-click reversible for the developer — so the audience is willing, not captive.
The audience: technical, in-flow, holding buying authority
The people watching that spinner are developers actively writing software with an AI agent. That's a tight, self-selecting slice of developer audience targeting: they use modern AI tooling, they're mid-task, and a large share of them choose or strongly influence the tools, infra, and APIs their team buys.
This is peak intent in the literal sense — they are building something right now, and the thing you sell (a database, an API, a dev tool, a framework) is the kind of thing they reach for next. A developer-focused ad campaign that lands here reaches programmers at the exact moment your category is relevant, not three days later in an unrelated feed.
We don't sell demographic profiles, retargeting cookies, or behavioral dossiers. The targeting is contextual: who's coding, in which editor, right now. Honest reach, not surveillance reach.
Format and pricing: 1,000-impression blocks, live bid
You buy in blocks of 1,000 impressions through a live bid market. The highest bid takes the #1 spinner slot, so pricing is set by real advertiser demand rather than a fixed rate card. As a current reference point, the top slot clears around a $20 CPM.
About half of every dollar you spend goes straight to the developer whose editor showed your line — roughly $10 reaches the developer per 1,000 impressions on that ~$20 CPM. Payouts settle in USDC on Solana mainnet and are verifiable on-chain via Solscan. This is real money moving on a public ledger, not credits in a dashboard.
There's no five-figure minimum and no invite list. You can buy your first 1,000 impressions and see real in-editor delivery, then scale your bid as the slot earns its keep.
- ▸Live bid market — highest bid wins the #1 spinner slot; ~$20 CPM as a current reference.
- ▸~$10 of a $20 CPM reaches the developer; paid in USDC on Solana mainnet, verifiable on Solscan.
- ▸Bought in 1,000-impression blocks, no large minimum, no invite required.
What's moderated and what's measured
Every line is a single piece of text, which keeps moderation tractable: no scripts, no autoplay, no tracking pixels riding along. We review creative for the obvious reasons — it has to be honest, relevant to a technical audience, and safe to show inside someone's working editor.
What's measured is deliberately narrow. The extension renders your ad and reports that a view happened. It never reads the developer's code, prompts, or files — privacy is enforced by the architecture, not by a promise in a policy. That means your reporting is impression-based and clean, and the developer's trust in the placement stays intact, which is what keeps them opted in and keeps your line in front of them.
Frequently asked questions
Who actually sees my ad?
Developers running an AI coding agent in VS Code, Claude Code, or Codex, while the agent is thinking. They've opted in and are paid for the view, so it's a willing, technical, in-flow audience — not a captive or tricked one.
How much does it cost to start?
You buy in blocks of 1,000 impressions in a live bid market, with no large minimum and no invite list. The #1 slot currently clears around a $20 CPM; you set your bid and can start with a single block.
How are developers paid, and is it real?
About half of your spend — roughly $10 per 1,000 impressions on a $20 CPM — goes to the developer in USDC on Solana mainnet. These are real on-chain payouts, verifiable on Solscan. It is not test money or dashboard credits.
Does claudearn track developers or read their code?
No. The extension renders your line and reports that a view happened — that's the entire list. It never reads code, prompts, or files. Your reporting is impression-based; the developer's privacy is enforced by the architecture.
What can I put in the line, and what gets rejected?
One short line of text — no scripts, autoplay, or pixels. We review every creative to make sure it's honest, relevant to a technical audience, and safe to show inside a working editor.