Fission-AI/OpenSpec

Spec-driven development (SDD) for AI coding assistants.

What it solves

OpenSpec addresses the unpredictability of AI coding assistants when requirements are stored only in chat history. It provides a lightweight specification layer that ensures humans and AI align on exactly what to build before any code is written, reducing vague prompts and unpredictable results.

How it works

OpenSpec integrates with over 30 AI assistants via slash commands (e.g., /opsx:propose) to create an artifact-guided workflow. It organizes development into a structured openspec/ directory containing plain Markdown files for proposals, requirements (specs), technical designs, and implementation checklists.

Key workflows include:

  • Explore: Using /opsx:explore as a thinking partner to weigh options and shape a plan.
  • Propose: Using /opsx:propose to generate a dedicated folder for a feature containing the proposal, specs, design, and tasks.
  • Apply: Implementing the defined tasks.
  • Archive: Moving completed changes to an archive to keep the active workspace clean.
  • Stores: A beta feature allowing teams to maintain specifications in a separate repository, creating a single source of truth shared across multiple code repositories.

Who it’s for

It is designed for solo developers and engineering teams who use AI coding assistants (such as Cursor, GitHub Copilot, Amazon Q, or Codex) and want a more predictable, organized, and scalable way to manage feature requirements and implementation.

Highlights

  • Tool Agnostic: Works with 30+ AI assistants through tailored slash commands.
  • Artifact-Guided: Uses plain Markdown for requirements and scenarios, making specs easy to read and review.
  • Cross-Repo Planning: Supports "Stores" to manage specifications for features that span multiple repositories.
  • Fluid Workflow: Avoids rigid phase gates, allowing iterative updates to any artifact at any time.

Related

  • Project
    clawplays/ospec

    A spec-driven, agentic workflow framework for AI coding agents that persists requirements, plans, and verification evidence directly in the project repository.

  • Project
    github/spec-kit

    An open-source toolkit that implements a spec-driven development process for AI coding agents, ensuring software is built from executable specifications rather than ad-hoc prompts.

  • Project
    lessweb/deepcode-cli

    A terminal-based AI coding assistant specifically optimized for DeepSeek models, featuring agent skills, MCP integration, and granular permission controls.

  • Dispatch
    OpenAI Model Spec

    OpenAI has introduced the Model Spec, a framework of objectives, rules, and default behaviors designed to standardize and transparently shape how AI models behave in the OpenAI API and ChatGPT.

  • Project
    OpenWhispr/openwhispr

    OpenWhispr is a cross‑platform desktop app for privacy‑first voice dictation, meeting transcription, and AI‑assistant interaction. It supports offline speech‑to‑text with Whisper‑based models or cloud services, offers speaker diarisation, note‑taking with semantic search, team collaboration, and a public API for developers.