Hashimoto argues that GitHub executed a bait-and-switch: offering free hosting to open-source projects, then monetizing their code through Copilot without meaningful consent or compensation. This is a core reason he is migrating Ghostty to a self-hosted forge, despite the significant cost of moving a fast-growing project with tens of thousands of stars.
The editorial argues that while forge migrations aren't new — GNOME moved to GitLab, the kernel uses mailing lists — Ghostty is different because it's a young, popular, contributor-heavy project led by one of the most credible names in open-source infrastructure. The willingness to absorb migration costs at this stage of growth indicates the depth of discontent has crossed a threshold that previously only foundations and smaller projects acted on.
By choosing to move Ghostty to a self-hosted forge rather than another centralized platform like GitLab or Codeberg, Hashimoto is implicitly endorsing self-hosting as a workable path for even large, active open-source projects. This follows the precedent set by projects like Forgejo and aligns with a broader trend of communities seeking infrastructure sovereignty.
The editorial highlights that Ghostty is a 'contributor-heavy project' that rapidly accumulated community engagement on GitHub, implicitly acknowledging the real cost of migration: leaving the platform where most developers already have accounts, where discovery happens organically, and where the PR/issue workflow is a shared lingua franca. The fact that this cost is being absorbed anyway underscores the strength of the opposing motivations.
Mitchell Hashimoto — co-founder of HashiCorp, creator of Terraform, Vagrant, and a half-dozen other tools that shaped modern DevOps — announced that Ghostty, his GPU-accelerated terminal emulator written in Zig, is leaving GitHub. The project will migrate to a self-hosted forge, joining a growing list of open-source projects that have decided GitHub is no longer the right home for their code.
Ghostty has been one of the breakout developer tools of the past year. After open-sourcing in late 2024, it rapidly accumulated tens of thousands of GitHub stars and built an active community of contributors. The fact that a project this young and this popular is willing to absorb the cost of a platform migration tells you something about how deep the discontent runs.
The announcement, posted on Hashimoto's personal blog, laid out a case that's been building across the open-source ecosystem for years — but rarely articulated by someone with this much credibility and this much to lose.
The forge migration trend isn't new. GNOME moved to GitLab years ago. The Linux kernel never left its mailing-list workflow. Codeberg and Forgejo have been gaining traction as community-run alternatives. But most migrations have been driven by organizations (foundations, enterprise projects) or by smaller projects where the switching cost is low. Ghostty is different because it's a fast-growing, contributor-heavy project led by one of the most recognized names in open-source infrastructure — and he's making the move anyway.
Hashimoto's argument hits on several pressure points that have been accumulating against GitHub:
The Copilot problem. GitHub Copilot was trained on the corpus of public repositories hosted on the platform. For many open-source maintainers, this felt like a bait-and-switch: host your code for free, then monetize it through an AI product without meaningful consent or compensation. The legal battles around this are ongoing, but the cultural damage is done. Maintainers who care about the ethics of their code's downstream use now have to actively think about whether GitHub is aligned with their values.
Platform lock-in disguised as convenience. GitHub Actions, GitHub Packages, GitHub Discussions, GitHub Projects — each feature tightens the dependency. What starts as "we'll just use Actions for CI" becomes a full-stack commitment that's genuinely painful to unwind. Hashimoto, who built an entire company around infrastructure tooling, knows platform lock-in when he sees it.
Philosophical alignment. There's a fundamental tension in hosting open-source projects on a proprietary platform owned by one of the largest software companies on earth. For years, the pragmatic argument won: GitHub is where the developers are, the network effects are too strong to ignore, and the alternatives weren't good enough. That calculus is shifting. Forgejo and Gitea have matured substantially. Self-hosting a forge in 2026 is dramatically easier than it was even three years ago.
Workflow degradation. This is the quieter complaint, but it resonates with anyone who's managed a large project on GitHub recently. Issue search is unreliable. The PR review interface hasn't meaningfully improved in years. Notifications are a mess. For a project like Ghostty that generates significant community activity, these aren't minor annoyances — they're daily friction that compounds.
The Hacker News discussion — scoring over 2,250 points — reveals a community that's deeply split but leaning sympathetic. The top comments reflect a recurring pattern: developers who *agree* with the critique but worry about the practical consequences. "I want to contribute to Ghostty, but I'm not creating an account on another forge" is a sentiment that appeared repeatedly. The migration tax on contributors is the single biggest counterargument, and it's legitimate.
If you maintain an open-source project, this is worth watching closely — not because you need to follow Ghostty off GitHub tomorrow, but because the ecosystem dynamics are shifting in ways that affect your decisions.
Discoverability matters more than you think. GitHub search, trending lists, and the social graph (stars, follows) drive a meaningful share of how developers find new tools. Moving off-platform means you need to replace that distribution channel. Ghostty can afford it because Hashimoto has a personal brand that transcends any single platform. Most projects can't.
CI/CD portability is an actual engineering problem. If your project relies on GitHub Actions with marketplace actions, custom runners, and matrix builds, migrating to another forge means rewriting your CI pipeline. This is the real lock-in — not the git repository itself, which is trivially portable, but the automation layer built on top of it. Projects considering a future migration should think about CI portability now, even if they're staying put.
Mirror strategies are becoming standard. Several projects that have left GitHub maintain a read-only mirror for discoverability while directing contributions to their primary forge. This is probably the pragmatic middle ground for most projects that share Hashimoto's concerns but can't absorb the full contributor friction of a complete departure.
For Ghostty users specifically: the terminal emulator itself won't change. Your config files, your keybindings, your workflow — all unaffected. You'll just need to update where you pull releases from and where you file issues. If you're building from source, you'll point your git remote at a new URL.
GitHub's dominance in open-source hosting has always rested on a network-effects moat: developers are there because other developers are there. Every high-profile departure chips away at that moat — not by draining it, but by proving the water on the other side is survivable. Ghostty won't trigger a mass exodus. But it gives every maintainer who's been privately frustrated with GitHub a concrete proof point that leaving is possible, even for a project with serious momentum. The question isn't whether more projects will follow. It's whether GitHub will respond with changes that address the underlying grievances, or whether it will keep betting that inertia is enough.
Top 10 dev stories every morning at 8am UTC. AI-curated. Retro terminal HTML email.