- Published on
OpenSpec β A lightweight and configurable AI spec framework
- Authors
- Name
- DeepHorus
OpenSpec, a lightweight and configurable specification framework from Fission-AI, has reached 68.0k stars on GitHub. The open-source project helps teams capture software requirements in specs and keep coding agents aligned as work evolves. According to its project page, a new spec is created with OpenSpec every two seconds. More than 265,000 developers use it monthly.
The framework addresses a common problem: requirements drift away from implementation. OpenSpec enables users to refine requirements, validate that they describe the right thing, and verify that the implementation matches. Its stated goal is to help developers build the right thing and build it right.
Key figures underline its rapid adoption. The repository holds 68.0k GitHub stars and is licensed under MIT. The current version is v1.13.0. Installation is available through npm, pnpm, bun, yarn, and nix, with the command npm install -g @fission-ai/openspec@latest.
OpenSpec organizes work through a command workflow. The /opsx: explore command maps the problem and codebase. /opsx: propose drafts proposal.md, specs/, design.md, and tasks.md. /opsx: apply implements tasks, /opsx: verify checks implementation against the spec, and /opsx: archive stores completed changes.
The tool is compatible with many AI coding environments. These include Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode, Amazon Q Developer, and dozens more. The directory lists over 33 additional integrations, showing broad ecosystem support.
OpenSpec is built in the open and used in the wild. Its community activity suggests that specification-driven development is becoming central to AI-assisted software work. The project's two-second spec interval demonstrates sustained, practical use.
With 265,000 monthly developers and 68.0k stars, OpenSpec has established a strong open-source presence. Its continued updates and integrations may further standardize how teams align humans and AI agents around shared specifications. The project remains open source under the MIT license.
