Open-source specs-driven development (SDD) framework for AI agent alignment. Bootstraps shared documentation and assistant guidance in minutes, creating a version-controlled context for both humans and AI agents.
Tagline: "Align AI agents, unify development teams"
AiCite is a powerful yet simple specs-driven development (SDD) framework that helps teams get started with SDD. It creates a shared context for both humans and AI agents by generating:
- Centralized documentation: Requirements, architecture, implementation, and deployment guides in
docs/ - AI agent guidance: Configuration for tools like GitHub Copilot, KiloCode, and Cursor IDE (with extensibility for more tools)
- Version-controlled context: All artifacts are local to your repository for full control
Stop wasting time aligning AI agents and human developers. AiCite provides a shared framework that ensures everyone works from the same page.
- Instant consistency: All AI agents (Copilot, KiloCode, Cursor) work from identical documentation
- Faster onboarding: New team members understand the project structure in minutes
- Reduced rework: Clear specs prevent misunderstandings before they happen
- Project tracking: Real-time progress updates through documentation status indicators
- Architecture-first mindset: Enforce design documentation before implementation begins
- Living specifications: Documentation evolves with your codebase
- Tool agnostic alignment: Ensure consistency across different AI platforms
- Version-controlled context: All specs and guidance are tracked in your repository
Get started with AiCite in under a minute!
npx aicite@latest setupuvx aicite setupnpx aicite@latest setupnpx aicite@latest setup --only docsnpx aicite@latest setup --copilotnpx aicite@latest setup --cursornpx aicite@latest setup --kilocodenpx aicite@latest setup --force--force: Overwrite existing generated files--only copilot,kilocode,cursor,docs: Generate only selected targets (docs are always included)--cursor/--copilot/--kilocode/--docs: Convenience flags for selective generation--cursor/--copilot/--kilocode/--docs: Convenience flags for selective generation
| Target | Description |
|---|---|
docs/ |
Requirements, architecture, implementation, and deployment guides |
copilot |
GitHub Copilot agent personas and guidance under .github/ |
kilocode |
KiloCode configuration including .kilocodemodes file and .kilocode/ folder |
cursor |
Cursor IDE agent configuration under .cursor/ |
| (future) | Support for additional AI tools and agents |
AiCite follows an SDD approach:
- Define requirements first: Clear, measurable objectives
- Architect before coding: Design solutions upfront
- Generate living docs: Specifications evolve with the project
- Align across tools: AI agents and humans work from the same source of truth
This ensures consistency, reduces rework, and improves collaboration between humans and AI.
AiCite's specs-driven approach enables powerful project tracking capabilities by maintaining up-to-date documentation with status indicators. AI agents can analyze these documents to provide real-time progress reports, identify blockers, and track dependencies.
Use these prompts to work with AI agents on architectural and project tracking tasks:
I need to define the requirements for a new feature that allows users to export their data. Help me update the requirements document.
We're planning to refactor our authentication system. Can you help me design the new architecture and document the changes?
I want to understand the current architecture of our project. Can you analyze the codebase and update the architecture document?
Generate a project tracking status report based on the current documentation.
What are the current project blockers based on the requirements and architecture documents?
Provide a summary of last week's status and current dependencies from the project documents.
Scan the project and prepare/update all documentation to reflect the current state.
Let's brainstorm solutions for the performance issues mentioned in the architecture document.
Use these prompts to work with AI agents on development and project tracking tasks:
I need to implement the user authentication feature. Can you help me understand the requirements and architecture, then guide me through the implementation?
There's a bug in the data export functionality. Can you help me debug it and fix the issue?
I'm refactoring the payment processing code. Can you review my changes and provide feedback on architectural alignment?
What is the high priority task to pick up next based on the requirements document?
Update the development progress status in the implementation document.
Let's see AiCite in action with a typical project scenario:
- Project initialization: A developer runs
npx aicite@latest setupto create the initial documentation and AI agent guidance. - Requirements gathering: The architect uses an AI agent with the prompt: "Help me define the requirements for a user authentication feature" and updates
docs/requirements.md. - Architecture design: The architect collaborates with an AI agent to design the authentication system and updates
docs/architecture.md. - Implementation: A developer uses an AI agent with the prompt: "Help me implement the user authentication feature based on the requirements and architecture" to write the code.
- Progress tracking: The team uses AI agents to generate status reports, identify blockers, and update documentation with real-time progress.
- Iteration: As the project evolves, the team updates the specifications, and AI agents provide consistent guidance across all tools.
This workflow ensures that everyone is aligned from the start, reduces rework, and improves collaboration between humans and AI.
AiCite solves the biggest challenges of AI-assisted development:
Ensure GitHub Copilot, KiloCode, Cursor, and future AI tools all work from the same specifications
Centralized documentation (requirements → architecture → implementation → deployment) that's version-controlled
Force best practices by documenting before coding, preventing costly rework later
Status indicators in documentation enable AI agents to provide real-time progress reports
One command setup (npx aicite@latest setup) with safe defaults that won't overwrite your existing work
Extensible framework that will support additional AI tools and workflows as the ecosystem evolves
We welcome contributions from the community! AiCite is built with specs-driven development, and we follow these principles in our own work. Here's how you can contribute:
- Fork the repository
- Clone your forked repository
- Set up the development environment (see
docs/implementation.mdfor details) - Make your changes
- Run the tests (see
docs/implementation.mdfor test commands) - Submit a pull request
- Follow the specs-driven development approach
- Keep changes focused on a single issue or feature
- Update documentation to reflect your changes
- Test your changes before submitting a pull request
- Be respectful and inclusive in your interactions with other contributors
If you find a bug or have a feature request, please open an issue on GitHub. Include as much information as possible, including steps to reproduce the issue.
AiCite is open-source software licensed under the MIT license.