A agentindex
Home · stack

Write code with AI

I want an AI that writes and edits code with me. Here's a stack that does this job, and a worked example of it in use.

Case study — Software developer

A developer pairs with an AI to ship a feature

A software developer is adding a feature to a web app. They want an AI that writes and edits code with them, not a search box that guesses.

Their stack has three pieces. Their codebase is what the agent works in. A coding agent reads the relevant files, proposes edits, and runs the tests; they might pick Cursor, an editor with the agent built in, or Claude Code, a terminal agent. Behind the agent is a model — Claude or GPT — which they either get bundled or pay for by token.

The workflow that works, drawn from long-running field notes on Claude Code, is to point the agent at the right files, let it make the change, and review every edit before it lands. The agent handles the typing and the boilerplate; the developer keeps the judgment.

Sources

  • Field notes from 2,000 hours with Claude Code (raw/0001-claude-code-2000h.md)
  • Cursor documentation (raw/0005-cursor-docs.md)

The stack

Click a box to filter the list to just the tools that fill it.

flowchart LR
  code(("Your code"))
  agent["Coding agent"]
  model["The model"]
  herd["Run several at once"]
  agent -->|reads and edits| code
  agent -->|sends prompts to| model
  herd -->|runs| agent

Coding agent

The AI that writes, edits, and navigates your code.

  • Aider — AI pair programming in your terminal, git-native.
  • Augment Code — A context engine for large enterprise codebases.
  • Claude Code — Anthropic's agentic CLI coding assistant.
  • Cline — Autonomous coding agent inside VS Code, with approval gates.
  • Continue — Open-source AI code assistant for VS Code and JetBrains.
  • Cursor — The AI-first code editor built on VS Code.
  • Devin — Cognition's autonomous software engineer.
  • Gemini CLI — Google's open-source terminal agent for Gemini.
  • GitHub Copilot — GitHub's AI pair programmer across editor and CLI.
  • Goose — Block's open-source local AI agent.
  • Hermes Agent — The self-improving AI agent built by Nous Research.
  • Kilo Code — Open-source coding agent consolidating Cline and Roo.
  • Open Interpreter — Let an LLM run code on your machine.
  • OpenAI Codex — OpenAI's open-source agentic coding CLI.
  • OpenCode — A model-agnostic, local-first coding agent TUI.
  • OpenCode Autolearn — Self-improvement engine for OpenCode agents.
  • OpenHands — Open-source autonomous software engineer (formerly OpenDevin).
  • Pi — A minimal, extensible terminal coding-agent harness.
  • Replit Agent — Build and ship apps conversationally on Replit.
  • Roo Code — A community fork of Cline with extra modes.
  • Tabby — Self-hosted AI coding assistant for teams.
  • Void — An open-source Cursor alternative.
  • Windsurf — Codeium's AI-native IDE (Cascade agent).
  • Zed (AI) — The high-performance editor with built-in AI.
See all on the directory →

The model

The brain behind the agent. You pick it, or bring your own key.

  • Anthropic — The Claude model API.
  • Cerebras Inference — Wafer-scale inference for very high throughput.
  • Cohere — Enterprise LLMs, embeddings, and rerankers.
  • DeepInfra — Low-cost hosted inference for open models.
  • Fireworks AI — Fast hosted open-model inference.
  • Google AI Studio — The Gemini API with a generous free tier.
  • Groq — Ultra-low-latency inference on custom LPU silicon.
  • llama.cpp — C/C++ inference for GGUF models on any hardware.
  • LMDeploy — Efficient inference and serving for open models.
  • Mistral AI — Provider of the Mistral and Codestral models.
  • Novita AI — Affordable hosted LLM and image APIs.
  • NVIDIA NIM — NVIDIA's inference microservices for hosting open and frontier models.
  • Ollama — Run large language models locally with one command.
  • OpenAI — The GPT and Codex model API.
  • OpenRouter — One API routing to hundreds of models across providers.
  • Replicate — Run and host thousands of open models via API.
  • SGLang — Fast serving and structured generation engine.
  • Text Generation Inference — Hugging Face's production LLM server.
  • Together AI — Hosted open-model inference and fine-tuning.
  • vLLM — High-throughput self-hosted LLM serving engine.
  • Z.ai — Provider of the GLM family of models.
See all on the directory →

Run several at once

A terminal multiplexer runs many agents in separate panes and keeps them alive when you disconnect.

  • Herdr — A terminal agent multiplexer that runs all your coding agents in one place.
  • tmux — A terminal multiplexer for managing many terminal sessions at once.
See all on the directory →