Penpot is the open-source design tool that renders in real web standards instead of a proprietary format, and its user base climbed from 250,000 in early 2023 to roughly 1.5 million by early 2026 (TechCrunch). A big chunk of that spike came the week Adobe announced it was buying Figma. Designers went looking for an exit, and Penpot was standing right there.
I've used Penpot on and off since the 1.x days, when it was promising but rough. The 2.0 line changed that. So is it finally a Figma replacement, or just a principled backup plan? Here's my honest read after real project work.
TL;DR: Penpot, on 2.17.2 as of late August 2026, is a genuine Figma alternative for product UI rather than a principled backup. It ships native design tokens (now edited in the Design tab, not a separate panel), CSS grid and flex layout, free developer handoff without a paid Dev seat, an MCP server, and real self-hosting. One gap is closing and one is not: rendering moved to WebGL and a WASM engine across 2.16 and 2.17, while the plugin ecosystem is still a fraction of Figma's. Most small-to-mid teams should pilot it now.
Is Penpot a real Figma alternative in 2026?
For product UI, yes, with eyes open about the gaps. Penpot now has the pieces a working team needs: components, flex and grid layout, native design tokens, and a plugin system. The catch is maturity in the corners. The plugin marketplace is a fraction of Figma's, and big shared libraries can feel sluggish. So it's a real alternative for focused product design and a weaker fit for a plugin-heavy, thousand-component monorepo. That's the trade, and it's a fair one for a tool you can self-host for free.
What has Penpot shipped since 2.0?
Penpot 2.0 shipped native CSS Grid layout alongside the existing flexbox, plus a full interface redesign and a reworked components system (Penpot). Why does that matter? Because both layout models map to what browsers actually do, so a grid you build in Penpot behaves like the grid a developer writes. There's no translation tax.
It also made Penpot the first design tool with native design tokens (Penpot). Not a plugin. Built in.
That was May 2024, and this article carried it as the current state for far too long. Here is what has actually shipped since, because two of these change the recommendation below rather than decorating it:
| Release | Date | What landed |
|---|---|---|
| 2.15.x | May to June 2026 | Plugin API token methods fixed on PRO, comment and font-name sanitisation, rate limiting on file snapshots |
| 2.16.0 | June 11, 2026 | WebGL rendering as a beta user preference, and design tokens moved into the Design tab itself, with numeric fields offering token selection in place |
| 2.16.x | June 2026 | The MCP server stopped being a beta signup: the MCP key now appears on the Integrations page, with Redis channel isolation for multi-tenant self-hosting |
| 2.17.0 | July 22, 2026 | The prototype viewer renders through a WASM Skia engine instead of SVG |
| 2.17.2 | August 27, 2026 | Current release. Typography-token and SVG-export fixes, plus a command-injection fix in the SVG exporter |
The MCP line matters most if you were here for the AI-workflow angle. In December 2025 it was a beta recruiting testers; by mid-2026 you generate a key in the product like any other integration. The rendering work matters for a different reason, covered under performance below.
How does Penpot handle design tokens and dev handoff?
Tokens sit at the center rather than bolted to the side. They connect to components and to grid and flex layouts, and a tokens API lets plugins read and write them. If you're building a system, that's the difference between one source of truth and a spreadsheet nobody trusts. It pairs naturally with the way design tokens reach real code.
Handoff is Penpot's quiet superpower. The CSS-native model means Inspect gives developers real CSS values and asset exports without a paid Dev seat in the way. Anyone with the file link reads production-shaped styles. For a small team, removing that paywall removes a recurring argument.
Penpot's free tier isn't a trial or a crippled demo. It's the actual product.
Penpot vs Figma: where each one wins
| Capability | Penpot | Figma |
|---|---|---|
| File format | Open, CSS-native | Proprietary |
| Free tier | Real product, 8 members | 3 files, limited |
| Dev handoff | Free Inspect, real CSS | Dev Mode (paid seat) |
| Self-hosting | Yes, unlimited, free | No |
| Design tokens | Native | Via Tokens Studio plugin |
| Plugin ecosystem | Small but growing | Large and mature |
| Large-file performance | Lags at scale, but 2.16 and 2.17 moved rendering to WebGL and WASM | Stronger |
The pattern is clear, with one row worth reading twice. Penpot wins on ownership, openness, and cost. Figma wins on ecosystem depth and, still, on raw performance at scale.
I want to be careful with that last row, because it is the claim most likely to be out of date by the time you read this. The 2.16 WebGL preference and the 2.17 WASM prototype viewer are precisely an attack on it, and the direction of travel is obvious. But a beta rendering preference is not the same as a solved problem, and I have not re-run a 500-component library against 2.17 myself. Treat "lags at scale" as true and shrinking, and test it on your own library before letting the row decide anything. If you want the full head-to-head against the other mainstream pick, our Figma vs Canva comparison covers where Figma fits, and the best UI design tools 2026 roundup puts all three side by side.
What does Penpot cost in 2026?
Less than you'd guess. Cloud Professional is $0 for up to 8 members with 10GB storage. Unlimited is $7 per user per month, capped at $175/mo total. Enterprise is $25 per user per month with a roughly $950/mo minimum (Penpot). Self-hosting is the real story: the Professional tier is free with unlimited users and files, so a 40-person team pays only for a server. Enterprise self-hosted adds SSO and SLAs at $950 per organization per month.
Who should switch, and who should wait?
Switch if you value owning your files, you self-host, or your budget can't absorb per-seat Figma pricing across designers and developers. A team standing up a fresh design system in Figma should at least pilot Penpot first, because the token layer is native here. Wait if you depend on specific Figma plugins, you run enormous shared libraries, or your prototyping needs are complex and runtime-heavy.
My take? Penpot crossed the line from "admirable underdog" to "defensible default for a lot of teams" somewhere in the 2.0 cycle. It isn't the smoother tool. It's the one that hands your work back to you, in a format the browser already understands, for free. That's worth a serious pilot before you renew another year of seats.
Neither editor sells hardware, so nothing here is part of the Penpot-versus-Figma call. It is the desk kit that sits under either choice: a pressure pen for the passes a mouse can't do cleanly, a programmable board for the shortcuts you run all day in whichever app you land on.
