Cline just launched a desktop app, but the real story is what it says about where AI pricing is headed for everyone, not just programmers.
Cline began in June 2024 when a solo developer named Saoud Rizwan built a weekend project for a hackathon hosted by Anthropic, the maker of Claude. That prototype, first called Claude Dev, went viral among programmers and grew into a company that has since raised 32 million dollars from investors including Emergence Capital. Today, more than 11 million developers use it, most of them through a free extension for the coding tool VS Code.
What made Cline stand out from day one was its pricing. Instead of charging a flat monthly fee like competitors Cursor and Claude Code, Cline charges nothing for the software itself and passes on the exact cost of whatever AI model you pick, with no markup added. That mattered because developers on flat-fee subscriptions kept hitting surprise limits: Cursor changed its pricing plan and left some customers with unexpected charges, while Anthropic quietly capped how much Claude Code power users could run each week. Cline built its pitch around avoiding exactly that kind of lock-in.
The new desktop app carries that pitch beyond a single code editor. You can now run several AI agents side by side, one drafting code, another researching a topic online, another handling a recurring task, all inside one workspace. You can schedule an agent to run automatically on a timer, the same way you would set a recurring meeting, so it checks for problems or writes a summary every morning without a person starting it each time. You can also pick up a half finished conversation from a rival tool like Claude Code or OpenAI's Codex and keep going in Cline with a different, cheaper model once your other subscription runs out of quota.
That last point matters more than it looks. Businesses paying for AI tools keep discovering the same thing: monthly quotas run out faster than expected, and the usual fix has been to pay for a bigger plan. Cline is betting people will instead spend less by switching to a cheaper, open model the moment the expensive one runs dry.
That bet is helped by a real shift already happening in the AI market. Open models built by Chinese labs, including DeepSeek and Moonshot's Kimi, now perform close to the level of expensive systems from OpenAI, Anthropic, and Google on real tasks, while charging a small fraction of the price per use. That gap has been narrowing fast, and it makes paying premium prices for AI harder to justify by the month.
This particular launch will not change coding work for most readers here. What it signals is broader: the choice between doing AI work through a costly, closed subscription or a cheap, open alternative is becoming a real decision rather than a default. Whoever controls that choice inside your business will decide how much you end up paying for AI from here on.