early access · developer network in waitlist rollout
// for advertisers
Reach engineers at the moment of creation.
While an AI coding agent works, its owner watches a single status line — focused, mid-build, undistracted. spinyield puts your message on it: one line, native to the tool, impossible to fake. We’re onboarding a small group of pilot advertisers now while the developer network rolls out from the waitlist.
⠧ Synthesizing… (esc to interrupt · 31s) · ✳ Your product, in one honest line — up to 60 chars
[01] the audience
The hardest audience in software, mid-task.
Engineers run ad blockers, skip pre-rolls, and don’t watch TV. But several times an hour, they sit and watch an AI agent work — eyes on one line of text, in the most focused context they occupy all day. That line is the inventory.
mid-build
your message lands while they're choosing tools, stacks, and services — not after the decision
native, not interruptive
one line in a surface they're already watching. No takeover, no autoplay, nothing to close
developer-priced attention
reach the people who pick the stack — the audience every devtool company pays a premium for
[02] proof-of-inference
The only inventory where bots run at a loss.
Everywhere else you buy ads, invalid traffic is filtered after the fact — partially, and you pay for the leakage. On spinyield, fraud is unprofitable before the fact.
every impression is backed by a paid inference
An ad renders only inside a live AI agent session — compute someone actually paid for. To fake one impression, a fraudster must burn more money on inference than the impression pays out.
cryptographically attested
Each impression is bound to a signed, single-use session token. Replays bounce, nonces never repeat, and sessions expire in minutes.
statistically defended
Frequency caps, cooldown gates, plausibility scoring, collusion detection, and payout holdbacks sit on top of the crypto. You're billed for verified moments, not estimates.
# anatomy of one billable impression
inference → agent session starts (real, paid compute)
session_token → signed, single-use, expires in minutes
view ≥ 5s → threshold met, event submitted once
fraud verdict → replay, caps, plausibility, collusion
ledger → you’re billed; 70% credits the developer
[03] economics
Metered, simple, and reviewed by humans.
pay per verified moment
Campaigns are priced per 1,000 verified impressions, set per campaign at pilot stage. A click bills at 50× an impression, capped per session — you never pay for a click storm.
one honest line
Creatives are a single line of up to ~60 characters plus a link. Every creative is reviewed by a human before it serves. No scams, no dark patterns — in either direction.
precise targeting
Target by agent (Claude Code, Codex), surface (editor or CLI), and version. Frequency caps are enforced per developer, so your line stays welcome.
01
tell us about your campaign
the form below — what you're promoting, rough budget, who it's for
02
we review and price it
human creative review, CPM quote, targeting setup. Within a business day
03
go live, get itemized reports
verified impressions and clicks, down to the event. Pause any time
[04] start
Put your line in the spinner.
We’re in waitlist rollout — onboarding a small group of pilot advertisers at founder pricing, with a direct line to the team. Reserve your spot now.
[05] faq
What buyers ask us.
Who actually sees my ad?
Developers using AI coding agents — Claude Code and Codex at launch — who have explicitly opted in to spinyield. They see it while their agent is working: the moment they're evaluating tools, infrastructure, and services.
How do I know the impressions are real?
Every impression is bound to a cryptographically signed session minted during a real, paid AI inference, then passed through replay checks, frequency caps, plausibility scoring, and collusion detection before it touches your invoice. Faking one costs a fraudster more than it pays — fraud here is unprofitable by construction, not filtered after the fact.
What does it cost?
Pilot campaigns are priced per 1,000 verified impressions, quoted when we review your campaign. Clicks bill at 50× an impression, capped per session. Early advertisers lock in pilot pricing.
What can the creative say?
One line, up to about 60 characters, plus your landing URL. Plain honest product messaging works best in this surface. Every creative is human-reviewed; we decline anything misleading, and developer trust is the asset we protect hardest.
Is this brand-safe?
Your line appears inside professional development tools, next to code and test output — there's no user-generated content to collide with. Brand safety runs both ways: we review creatives so developers trust the surface your ad lives in.
How big is the audience right now?
Early. The developer extension is in final development with a growing waitlist, and the serving backend is already live in production. Pilot advertisers come aboard now to lock pricing and shape targeting before general availability — we'll be straight with you about volumes.