A self-hosted personal AI assistant that runs on your own devices and answers you through messaging apps like WhatsApp, Telegram, Slack and Discord.
Fission-AI/OpenSpec
A spec-driven development workflow for AI coding assistants that agree on proposals, specs and tasks before any code is written.
Open-source alternative to:GitHub Spec Kit
Kiro
- Stars
- 72,380
- Forks
- 4,947
- Open issues
- 197
- Last commit
- yesterday
What is it?
OpenSpec is a lightweight spec layer for AI coding assistants. Each change gets its own folder containing a proposal, requirement specs with scenarios, a design document and a task checklist, so the human and the assistant agree on what to build before implementation starts.
The workflow is driven by slash commands such as /opsx:explore, /opsx:propose, /opsx:apply and /opsx:archive. Specs are plain Markdown, and artifacts can be updated at any point rather than passing through rigid phase gates. Changes are archived into a specs folder once implemented, so the project's specs stay current.
OpenSpec is installed via npm or Homebrew and requires Node.js 20.19.0 or later. It supports more than 30 AI tools and also offers Stores, a beta feature for keeping specs in a separate repository shared across several code repos.
Key features
- 01Per-change folders with proposal, specs, design and tasks
- 02Slash commands for exploring, proposing, applying and archiving changes
- 03Plain Markdown specs with WHEN/THEN scenarios
- 04Works with 30+ AI coding assistants
- 05Iterate on any artifact without fixed phase gates
Use cases
- Agreeing on requirements with an AI assistant before it writes code
- Adding features to an existing (brownfield) codebase with a reviewed plan
- Sharing a single set of specs across multiple repositories in a team
Quick start
- 1
Install Node.js 20.19.0 or higher (Homebrew installs it as a dependency)
- 2
Install OpenSpec globally with npm
npm install -g @fission-ai/openspec@latest - 3
Navigate to your project directory and initialize
cd your-project openspec init - 4
In your AI assistant, start with /opsx:explore or go straight to /opsx:propose <what-you-want-to-build>
Things to know
- Requires Node.js 20.19.0 or higher.
- Stores, a team-sharing feature, are in beta.
- Works best with high-reasoning models, which suggests extra cost for model access.
Topics
FAQ
- What is OpenSpec?
- A spec-driven development workflow for AI coding assistants that agree on proposals, specs and tasks before any code is written. OpenSpec is a lightweight spec layer for AI coding assistants. Each change gets its own folder containing a proposal, requirement specs with scenarios, a design document and a task checklist, so the human and the assistant agree on what to build before implementation starts.
- Is OpenSpec free to use? What license does it have?
- OpenSpec is open source under the MIT license. Check the license terms before using it in commercial products.
- Is OpenSpec still maintained?
- Actively maintained. The last commit to OpenSpec was on Oct 19, 2026.
- What can I use OpenSpec for?
- Agreeing on requirements with an AI assistant before it writes code Adding features to an existing (brownfield) codebase with a reviewed plan Sharing a single set of specs across multiple repositories in a team
- How do I install OpenSpec?
- Install Node.js 20.19.0 or higher (Homebrew installs it as a dependency) Install OpenSpec globally with npm: npm install -g @fission-ai/openspec@latest Navigate to your project directory and initialize: cd your-project openspec init In your AI assistant, start with /opsx:explore or go straight to /opsx:propose <what-you-want-to-build>
- Where can I get OpenSpec?
- The source code and installation instructions are on GitHub: https://github.com/Fission-AI/OpenSpec
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A collection of agent skills for coding assistants like Claude Code and Codex, covering grilling sessions, TDD, bug diagnosis, planning and code review.
An MCP server that gives coding agents IDE-level symbol search, semantic editing and refactoring over a codebase.
Turns a Git repository or local directory into a single prompt-friendly text digest for feeding code to LLMs.
An example Claude Code project setup showing CLAUDE.md, hooks, skills, agents, slash commands, MCP servers and GitHub Actions workflows you can adapt.
A collection of Claude Code skills from HumanLayer for research, design, PR review visuals, and agentic coding workflows.