A curated list of tools, frameworks, and resources for AI-powered software engineering.
Modern software development is being transformed by AI agents, coding copilots, and spec-driven workflows. This repository collects the best resources for developers who want to leverage AI as a true engineering partner.
- Philosophy
- Quick Start
- AI Coding Agents
- Coding Copilots
- Multi-Agent Frameworks
- Agent SDKs
- LLM Observability & Gateways
- LLM Inference Engines
- Spec-Driven Development
- Skills & Plugins
- Methodologies
- Learning Resources
AI-powered software engineering isn't about replacing developers—it's about amplifying their capabilities. The most effective approach combines:
- Specifications before code - Clear intent leads to better outcomes
- AI as collaborator - Leverage AI for execution while you focus on architecture
- Living documentation - Keep context fresh and accessible
- Iterative refinement - Build in feedback loops
Read more in our Philosophy Guide.
Step 1: Choose your primary tool
- New to AI coding? Start with Claude Code or Cursor
- Want maximum control? Try Aider or OpenCode
- Building custom agents? Use Claude Agent SDK
Step 2: Adopt a spec-driven workflow
- Write a brief PRD before starting any feature
- Use the PRD template to structure your requirements
Step 3: Iterate and refine
- Review AI-generated code critically
- Build your prompt library for common patterns
- Track what works in your CONTEXT.md
See our full Getting Started Guide.
Autonomous agents that can plan, write, and debug code with minimal supervision.
| Name | Author | Description |
|---|---|---|
| Claude Code | Anthropic | Official CLI for Claude that handles complex software engineering tasks autonomously |
| Codex CLI | OpenAI | Terminal-based coding agent powered by GPT models |
| OpenCode | Anomaly | Open-source AI coding agent with extensible architecture |
| Aider | Paul Gauthier | AI pair programming in your terminal with git integration |
| cecli (aider-ce) | dwash96 | Aider fork with a different style of AI-assisted coding |
| Mentat | Abante AI | AI coding assistant that understands your entire codebase |
| GPT Engineer | GPT Engineer | Generate entire codebases from a single prompt |
| Sweep | Sweep AI | AI-powered junior developer that handles GitHub issues |
| Mistral Vibe | Mistral AI | CLI coding assistant with interactive chat, code search, and shell integration |
| Google Antigravity | Agentic IDE with multi-agent deployment powered by Gemini 3 | |
| OpenAI Codex | OpenAI | Cloud-based coding agent for autonomous development tasks |
| GitHub Copilot CLI | GitHub | Terminal-based AI assistant with agent delegation and parallel execution |
| GitHub Copilot Coding Agent | GitHub | Cloud-hosted autonomous agent that works on GitHub issues |
IDE-integrated assistants that augment your coding workflow.
| Name | Author | Description |
|---|---|---|
| Cursor | Cursor Inc | AI-first code editor built on VSCode with deep model integration |
| Cline | Cline | Autonomous coding agent that lives in your IDE |
| Kilo Code | Kilo Org | VSCode extension for AI-assisted development with MCP support |
| GitHub Copilot | GitHub | AI pair programmer suite with Chat, CLI, coding agent, and code review agent |
| Continue | Continue | Open-source autopilot for software development |
| Supermaven | Supermaven | Ultra-fast code completions with 1M token context |
| Tabnine | Tabnine | AI code assistant trained on your codebase |
| Windsurf | Codeium | AI-first IDE with Cascade agentic flow and multi-file editing |
| Roo Code | Roo Code Inc | Open-source AI assistant with role-specific modes (Architect, Code, Debug, Test) |
Coordinate multiple AI agents for complex engineering tasks.
| Name | Author | Description |
|---|---|---|
| Gas Town | Steve Yegge | Multi-agent coding framework with specialized agent roles |
| Ralph | Snarktank | PRD-driven development with automated planning loops |
| CrewAI | Joao Moura | Framework for orchestrating role-playing autonomous agents |
| AutoGen | Microsoft | Multi-agent conversation framework ( |
| MetaGPT | MetaGPT | Multi-agent framework that turns requirements into PRDs and code |
| ChatDev | OpenBMB | Collaborative AI agents simulating a software company |
Build production AI agents programmatically with these SDKs.
| Name | Author | Languages | Description |
|---|---|---|---|
| Claude Agent SDK | Anthropic | Python, TypeScript | Build production agents using Claude Code as a library. Built-in tools, hooks, subagents, MCP support |
| Microsoft Agent Framework | Microsoft | Python, .NET | Unified framework combining AutoGen and Semantic Kernel (GA Q1 2026) |
| LangChain | LangChain | Python, JS | Popular framework for LLM-powered applications |
| LlamaIndex | LlamaIndex | Python | Data framework for LLM applications with RAG |
Monitor, debug, and route LLM traffic in production.
| Name | Author | Type | Description |
|---|---|---|---|
| LiteLLM | BerriAI | Gateway | Unified API proxy for 100+ LLM providers with load balancing, spend tracking, and caching |
| Langfuse | Langfuse | Observability | Open-source LLM engineering platform with tracing, prompt management, and evaluations |
| Helicone | Helicone | Observability | LLM observability with request logging, cost tracking, and prompt analytics |
| OpenLLMetry | Traceloop | Observability | OpenTelemetry-based tracing for LLMs, integrates with Datadog, Honeycomb, Grafana |
| Logfire | Pydantic | Observability | Python-native observability with OpenTelemetry, excellent FastAPI/Pydantic integration |
Self-host and run LLMs on your own infrastructure.
| Name | Author | Description |
|---|---|---|
| vLLM | vLLM Project | High-throughput LLM serving with PagedAttention for efficient memory management |
| llama.cpp | GGML | C/C++ LLM inference with broad hardware support (CPU, GPU, Apple Silicon) |
| SGLang | SGLang Project | Fast serving framework with RadixAttention and structured generation support |
Tools and methodologies for requirements-first AI development.
| Name | Author | Description |
|---|---|---|
| BMAD Method | BMAD | Breakthrough Method for AI-Driven development with structured specs |
| Shotgun | Shotgun | Spec-based code generation with planning phase |
| Spec-Kit | GitHub | Toolkit for managing specifications in AI workflows |
| OpenSpec | Fission AI | Open standard for AI-consumable specifications |
| Get Shit Done | Glitter Cowboy | Opinionated workflow for shipping with AI assistance |
| SpecStory | SpecStory | Turn user stories into detailed specifications |
| nano-spec | tao-hpu | Lightweight task specification framework with 4-doc structure for AI development |
Extend your AI coding tools with custom capabilities.
| Name | Author | Description |
|---|---|---|
| awesome-claude-skills | ComposioHQ | Curated collection of Claude Code skills and extensions |
| everything-claude-code | Affaan M | Comprehensive guide and resources for Claude Code |
| awesome-claude-code | hesreallyhim | Collection of Claude Code tips, tricks, and skills |
| awesome-claude-skills | VoltAgent | Skills library for extending Claude Code capabilities |
| claudekit-skills | mrgoonie | Ready-to-use skills for Claude Code workflows |
| MCP Servers | Anthropic | Official Model Context Protocol server implementations |
Workflows and approaches for AI-assisted development.
The PRD (Product Requirements Document) loop is a cyclic workflow:
- Specify - Write clear requirements in a PRD
- Plan - Let AI break down the work into tasks
- Execute - AI implements while you review
- Refine - Update the PRD based on learnings
- Repeat - Continue until complete
Start every feature with a specification:
- Forces clarity of thought before coding
- Gives AI maximum context to work with
- Creates documentation as a byproduct
- Enables better code review
Use multiple specialized agents for complex projects:
- Architect Agent - Designs system structure
- Developer Agent - Writes implementation code
- Reviewer Agent - Checks code quality
- Tester Agent - Creates and runs tests
See our Spec-Driven Development Guide for details.
| Guide | Description |
|---|---|
| Getting Started | Your first steps with AI-powered development |
| Spec-Driven Development | Deep dive into specification-first workflows |
| Choosing Your Tools | How to select the right AI coding tools |
| Philosophy | Core principles and mindset |
| Glossary | Key terms and definitions |
Ready-to-use templates for your projects:
- PRD Template - Product requirements document
- AGENTS.md Template - Multi-agent configuration
- CONTEXT.md Template - Project state tracking
- Anthropic Claude Documentation
- Claude Agent SDK Documentation
- Model Context Protocol
- Prompt Engineering Guide
We welcome contributions! Please see our Contributing Guidelines for details.
- Ensure the resource is actively maintained
- Verify it adds unique value to the list
- Submit a PR with the appropriate category
- Include a clear, concise description
MIT - Feel free to use this for your own awesome lists.
This list is inspired by the amazing work of:
- The awesome list community
- All the developers building AI coding tools
- Everyone sharing their workflows and learnings
Built with AI assistance, curated by humans.
