Seventy-two hours. That's all it took engineer Igor Šćekić to build, test, and launch Kilo Pass, the AI coding company's subscription service that went live in mid-January. While enterprise software teams typically measure product cycles in quarters, Kilo's approach was different—perhaps recklessly so, or perhaps shrewd. Šćekić shipped the initial version over a long weekend, then spent five more days iterating based on user feedback before calling it done.
The timeline sounds improbable, at least by conventional software development standards. It also reveals something essential about how Kilo, a startup barely past its seed round, believes it can compete against GitHub Copilot, Cursor, and Anthropic's Claude—incumbents with vastly deeper pockets and brand recognition.
Speed, in Kilo's calculus, isn't just an operational advantage. It's the entire strategy.
The company raised $8 million in December 2025, led by Cota Capital, a respectable but unremarkable seed round in an AI market drowning in capital. By the time Kilo Pass launched weeks later, the platform was claiming 1 million developers and 20 trillion tokens processed—impressive figures that the company self-reported, though it hasn't offered independent verification or specified how it defines an active developer versus someone who signed up and never returned.
The Economics: Straightforward to a Fault
Kilo Pass doesn't bury its pricing in fine print or obscure the math. The subscription converts dollars to credits at a 1:1 ratio, matching provider API rates across more than 500 models with zero markup. Pay $49, get $49 in credits. Those credits work across Kilo's entire ecosystem: VS Code extension, JetBrains IDEs, command-line interface, Cloud Agents, the Gateway API.
Three tiers exist: Starter at $19 monthly, Pro at $49, Expert at $199. What makes the subscription more than simple prepayment is the bonus structure, which escalates over time. Monthly subscribers receive a 50% welcome bonus in month one, then earn increasing bonuses that climb to 40% by month eight. Annual subscribers skip the ramp entirely and collect a 50% bonus every month from day one. Under that math, the Expert tier delivers up to $278.60 in total monthly credits after bonuses.
Paid credits never expire—an important detail. Bonus credits, however, only unlock after you've burned through that month's paid allocation, and they vanish at month's end if left unused. It's a system designed to reward commitment without penalizing intermittent use, though the escalating bonus tiers also function as a retention hook.
What Competitors Got Wrong, According to Kilo
CEO Scott Breitenother's explanation for why Kilo Pass exists doubles as a critique of the broader AI coding subscription market. In his January launch post, he argued that industry players had built their pricing models on an assumption that inference costs would plummet. When token usage surged and prices stayed stubbornly high, competitors responded with throttling and what Breitenother called "opaque credits"—systems where users couldn't easily track what they were actually paying per token.
Kilo's answer, as Breitenother described in his launch post, was transparency bordering on radical simplicity. The company keeps its pay-as-you-go option available while positioning the subscription as a "momentum subscription," a term Breitenother used to describe a product designed to keep developers in flow state rather than function purely as a discount vehicle.
Documentation explicitly states that $1 becomes $1 of Kilo credits. The pricing page lists "No markup" and "No rate limits" in bold. Teams and Enterprise plans, priced at $15 per user monthly and custom rates respectively, don't include model credits—a potential point of friction for organizations expecting an all-inclusive package.
Whether developers care about this level of transparency remains an open question. Convenience often trumps principle in software purchasing decisions.
Engineers as Mini-CEOs
A Fortune profile published in March offered a glimpse into Kilo's internal structure, which appears designed—intentionally or by necessity—to eliminate bottlenecks. As the Fortune feature described, the company employs engineers not as code executors but as "mini CEOs," owning features end-to-end from conception through deployment. At the time of the interview, Kilo had 14 engineers and was shipping more than 15 features weekly, a pace that would be unsustainable for most teams.
Kilo Pass itself was three days old when Šćekić built it. That timeline would be impossible under traditional product management hierarchies, where feature requests typically route through planning committees, design sprints, and stakeholder alignment meetings. Kilo appears to have dispensed with most of that apparatus.
Cofounders Scott Breitenother and Sid Sijbrandij—the latter GitLab's former CEO, a man who presumably understands enterprise software development at scale—seem to have bet that organizational leanness matters more than process rigor, at least at this stage. When users requested a subscription option, an engineer simply built it. No roadmap debates, no quarter-long timelines.
The approach carries obvious risks. Shipping fast can mean shipping broken, and technical debt accumulates quickly when there's no one playing defense. But for a seed-stage company trying to establish traction against entrenched competitors, moving slowly is its own form of risk.
Adoption Metrics and Product Expansion
Kilo's public claims about usage have escalated quickly. The January announcement cited 1 million developers. By the time the company's website was last updated—timing unclear—the Launch Partner page claimed 2.3 million downloads, a figure reported without clarification of whether those represent unique users, total installs across multiple platforms, or some other metric. Treated with appropriate skepticism, the numbers suggest momentum, even if the definitions remain fuzzy.
Kilo Pass credits now work across an expanded product suite. The company shipped a rebuilt VS Code extension in early April, unified with its CLI and Cloud offerings via an OpenCode server core. KiloClaw, Kilo's deployment platform for OpenClaw agents, reached general availability sometime between February and March with pricing that allows Kilo Pass subscriptions to cover both hosting (listed at $9 monthly) and inference costs.
A promotional "Double Welcome Bonus" extending the first-month 50% bonus to two months was scheduled to end April 10, suggesting the company is still experimenting with acquisition incentives.
The Harder Questions
What Kilo is testing, ultimately, is whether the kind of radical transparency and breakneck execution velocity the company champions can carve out defensible market share in a category where switching costs are low and user loyalty is mostly theoretical. The three-day build timeline for Kilo Pass makes for good storytelling. Whether it translates into sustainable growth is another matter.
The company hasn't disclosed renewal rates, churn figures, or revenue. Those are the metrics that will determine whether Kilo's speed-first philosophy actually works or simply produces a pile of features no one uses long-term.
For now, Kilo seems to be betting that developers will reward directness and speed—that a subscription with transparent economics and no artificial limits is compelling enough to pull users away from more polished, better-funded alternatives. It's not an unreasonable bet. It's also not a guaranteed one.
The subscription launched three days after someone decided it should exist. The real test begins now, in the months ahead, when the novelty fades and retention becomes the only number that matters.
