A new code editor called Gram has surfaced as a streamlined alternative to the popular Zed editor. As reported by Hacker News, Gram is a fork of Zed that deliberately strips away artificial intelligence features to focus purely on performance and configuration. This launch represents a pushback against the increasing integration of LLMs into developer environments, offering a tool for those who want a quiet, deterministic workspace.
This release highlights a specific niche in the developer community: users who admire Zed’s architecture but dislike the recent pivot toward AI assistants. Gram retains the underlying speed that made the original editor famous while removing the telemetry and cognitive load associated with AI features.
What Gram brings to the table
Gram positions itself as a “batteries included” editor that works immediately upon installation. Key features include:
- Core Performance: It maintains the high-performance rendering and responsiveness characteristic of the Zed codebase.
- Extension Compatibility: While it removes the AI, it keeps the utility. Gram supports many popular languages natively and can use existing Zed extensions to fill in gaps.
- Integrated Tooling: The editor ships with built-in documentation, Git source control, and debugger support via the Debug Adapter Protocol (DAP).
- Configurability: It is designed to be highly configurable, allowing developers to tune the environment to their specific workflow needs.
Why a non-AI fork matters
The tagline for the project, a code editor for “humanoid apes and grumpy toads,” signals a playful but clear philosophy. By quoting Ursula K. Le Guin (“What cannot be mended must be transcended”), the creators suggest that rather than disabling AI toggles in existing tools, a clean break was necessary.
For developers, this means access to a modern, fast editor that doesn’t nudge them toward generative code or cloud-based inference. It serves as a robust option for those working in air-gapped environments, on privacy-sensitive projects, or simply for those who prefer to write every line of code themselves.
You can find more details on the project’s repository linked via the original Hacker News discussion.