Skip to content

kicchann/reqord

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

770 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Reqord

Reqord

日本語

Structured context for the AI era, version-controlled alongside your code.

npm License Tests TypeScript


Reqord stores requirements as structured data (YAML + Markdown) inside your Git repository, giving every requirement a clear lifecycle from draft to implementation. No SaaS, no backend -- just git clone and your entire requirements history is there, version-controlled alongside your code, ready for humans and AI tools alike.

Reqord Dashboard — Project health score, requirement/specification progress, and warnings at a glance


Why Reqord?

Software teams lose track of requirements. The consequences compound over time:

  • Scattered requirements -- decisions live in chat threads, meeting notes, and memory. Nothing is canonical.
  • Lost context -- new team members and AI tools start every session without project background.
  • Spec drift -- specifications written at project start are never updated. Implementation quietly diverges.
  • Broken traceability -- no one can answer "why does this feature exist?" or "what breaks if we change it?"
  • AI gets bad inputs -- unstructured text produces inconsistent AI output. Every session starts from scratch.

Reqord solves these by making requirements structured, versioned, and traceable -- stored right where your code lives.

How It Works

Reqord enforces a 3-layer traceability model that connects intent to implementation:

Requirement (What)  -->  Specification (How)  -->  GitHub Issue (Tasks)
       ^                                                  |
       '------------------ Feedback Loop -----------------'
Layer Purpose Example
Requirement What to build "Users can log in with email"
Specification How to build it "OAuth2 + JWT, sessions stored in Redis"
GitHub Issue Concrete implementation tasks "Implement POST /auth/login endpoint"

Each layer links to the others. When a requirement changes, you can trace the impact through specifications to issues. When implementation feedback surfaces, it flows back to update requirements.

Lifecycle:

draft --> pending_approval --> approved --> implemented --> deprecated

Requirements move through a defined lifecycle with PR-based approval gates, so nothing gets lost and nothing ships without review.

Quick Start

Install

npm install -g @reqord/cli

# Optional: install the web dashboard
npm install -g @reqord/web

Claude Code Plugin

Reqord provides a Claude Code plugin for AI-assisted requirements workflow (design, TDD implementation, review, Git operations).

# Add the marketplace
/plugin marketplace add kicchann/reqord

# Install the plugin
/plugin install reqord@reqord-plugins

Initialize a project

cd /path/to/your/project

# Initialize the .reqord/ directory structure
reqord init

# Set up project context
reqord context init

Create your first requirement

# Create a requirement (supports user-story, ears, and free-form formats)
reqord req create

# List all requirements
reqord req list

# Validate requirement quality with SMART scoring
reqord req validate req-000001

See the Getting Started guide for the full walkthrough.

Development setup (building from source)

Prerequisites: Node.js 20+, pnpm 10+, Git

git clone https://github.com/kicchann/reqord.git
cd reqord
pnpm install
pnpm build
cd packages/cli && pnpm link --global

Key Features

Hybrid Storage (YAML + Markdown)

Requirements are stored as YAML metadata (status, priority, dependencies, version history) paired with Markdown content (descriptions, success criteria, use cases). Machine-readable structure with human-readable documentation.

SMART Validation

Built-in quality scoring based on the SMART framework (Specific, Measurable, Achievable, Relevant, Time-bound). Run reqord req validate to get an objective quality score and actionable improvement suggestions for any requirement.

PR-Based Approval Workflow

Requirements follow the same review process as code. Submit a requirement for approval, and it creates a pull request. CODEOWNERS review and merge to approve. The entire approval history is tracked with version, commit hash, reviewer, and timestamp.

Requirement Formats

Supports industry-standard formats out of the box:

  • EARS (Easy Approach to Requirements Syntax) -- structured trigger/condition/action format
  • User Story -- "As a [role], I want [feature], so that [benefit]"
  • Free-form -- flexible format for early-stage requirements

Web Dashboard

Launch the built-in web dashboard to visualize project health, dependency graphs, and requirement details:

reqord ui

Dependency graph — requirement-to-specification-to-issue relationships

AI Integration via ProjectContext

The .reqord/context/ directory stores structured project knowledge that AI tools can consume directly:

File Purpose
product.yaml Vision, target users, core features, out-of-scope
technical.yaml Tech stack, architecture, design patterns
structure.yaml Naming conventions, directory structure, import rules
domain/*.md Domain-specific knowledge (security policies, API standards)

This context persists across AI sessions, eliminating the "explain the project from scratch every time" problem. Works with Claude Code, Cursor, Windsurf, Codex, and any tool that can read files.

Feedback Loop

Track feedback from GitHub Issues back to requirements and specifications:

reqord feedback sync    # Sync feedback from GitHub Issues
reqord feedback list    # View feedback items
reqord feedback link feedback-001  # Link feedback to a requirement

When implementation reveals that a requirement needs updating, the feedback loop ensures that knowledge flows back upstream rather than getting lost in issue comments.

Comparison

How Reqord compares to existing tools for managing what to build:

Capability Reqord Jira Linear Notion GitHub Projects
Git-native Yes -- lives in your repo No No No Partial
Offline-first Yes -- no server needed No No No No
AI-ready structure YAML + Markdown, typed schemas Unstructured text Unstructured text Unstructured text Unstructured text
Traceability Enforced (Req -> Spec -> Issue) Manual linking Manual linking Manual linking Manual linking
Approval workflow PR-based with CODEOWNERS Built-in workflow Built-in workflow No formal approval No formal approval
Version control Full Git history + semantic versions Limited history Limited history Page history No versioning
Quality validation SMART scoring built-in No No No No
Cost Free (AGPL-3.0) Paid Paid Freemium Free (limited)

Reqord is not a replacement for project management tools. It focuses specifically on requirements lifecycle management and is designed to complement tools like Jira, Linear, or GitHub Projects.

Architecture

Reqord is a monorepo with three packages:

Documentation

See docs/README.md for the full documentation index.

Roadmap

Reqord is at v0.2.0. Features shipped since v0.1.0:

  • PR-based approval workflow -- reqord req approve / reqord spec approve with GitHub PR integration
  • Impact analysis -- reqord impact analyze / reqord impact notify for tracing change propagation
  • Task generation -- reqord task create / reqord task sync for decomposing specs into GitHub Issues
  • Status dashboard -- reqord status for project-wide progress overview
  • Project Settings -- setting.yaml for per-project CLI configuration

Upcoming:

  • Web UI enhancements -- Gantt charts, richer dependency visualization, specification editing

See the CLI Reference for the full list of commands and their current implementation status.

Contributing

Contributions are welcome. Please see CONTRIBUTING.md for guidelines.

License

AGPL-3.0

Reqord is licensed under the GNU Affero General Public License v3.0. You can freely use, modify, and distribute Reqord for any purpose, including commercial use. If you run a modified version as a network service, you must make the source code available to users of that service.

About

Git-native requirements management for the AI coding era. Structured YAML + Markdown, version-controlled alongside your code.

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages