The pressure valve

Figma↗ Desktop 126.1.2 shipped on February 27 with a change so small Figma↗ did not bother announcing: the --remote-debugging-port flag was gone. No deprecation notice, no migration path. The flag was how third-party tools had been talking to Figma↗ from outside — design automation scripts, AI agents, MCP servers. After the update, they stopped working.

The same day, a developer who had also had a Figma↗ plugin rejected without explanation pushed the first commit to OpenPencil↗, an open-source AI-native design editor, to GitHub under an MIT licence.

Within days it had 2,224 stars. A single tweet from @nunowar — "A Figma↗ plugin developer had their plugin silently rejected by Figma↗ and decided to build an opensource version of Figma↗ instead" — got 238,000 views.

The speed of the response says something. The specific grievance matters less than the fact that one afternoon of frustration produced a credible alternative, and one week later it went viral.

Figma↗ has been under different pressure since its IPO in July 2025. The stock opened at $85, peaked at $143, and was trading at $19.85 by February — a fall of 85% from peak — with a net loss that widened to $1.25 billion despite $1.06 billion in revenue. The path to profitability from that position involves familiar moves: tighten professional plan pricing, add Dev Mode charges, close the platform to third-party automation Figma↗ does not control. This is not unusual. It is what public companies do when growth is not enough.

Open-source design tools have been around since GIMP↗ in 1996 and Inkscape↗ in 2003. They have consistently been powerful for the right workflows and consistently failed to displace commercial platforms for professional use. The reason is not principle — designers are not opposed to open source ideologically. The reason is that real-time collaboration, performance at scale, and the polish that makes a tool feel fast all cost money to build. Figma↗ raised more than $200 million and spent it winning that race. The open-source alternatives were good but never good enough for teams who needed a single source of truth across twelve designers in four time zones.

Penpot↗ changed some of that. It is SVG-native, self-hostable, genuinely capable for teams who want to own their data. Getting there took years of serious engineering, and it still lags on integrations.

OpenPencil↗ is pre-1.0 in ways that matter for production work and interesting in ways that matter for the argument. It reads and writes .fig files natively. It runs locally with no account required. It ships with an MCP server so Claude Code↗ and Cursor↗ can talk to the canvas directly. The technical ambition is serious; it launched three weeks ago.

I do not think OpenPencil↗ is going to replace Figma↗. What it does is apply pressure at exactly the moment Figma↗ is squeezing its developer community, and that changes the negotiation.

When Figma↗ removed that debugging flag, they made a calculation: the cost to third-party developers is acceptable. That calculation shifts if those developers have somewhere credible to point to. "I will just move to OpenPencil↗" does not have to be true to be useful. It just has to be plausible, and for the first time in a while, it is.

What AI has done to the economics here is real. Getting a local-first vector editor with decent code export to a credible beta used to require a company-scale investment. Now one developer with a specific grievance can ship a working prototype in days and get 2,000 stars before the week is out. The floor for what counts as a pressure valve has dropped considerably.

Whether OpenPencil↗ matures into something teams actually use for production work, I genuinely do not know. What the remote debugging port decision tells you is something about where Figma↗'s priorities are pointing. It is a small change. That is what makes it worth paying attention to.