Building with AI

AI Coding Assistants

Use an AI assistant to write, edit and explain code in your own project, from autocomplete to agents that carry out multi-step tasks.

Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

The real choice

Where the agent works. AI-first editors put completions, chat and agents inside the editor where you watch every change; terminal agents take a task and work across the repo, running commands, with you reviewing the result; editor plugins add AI to the editor you already use without switching; agent managers run several agents in parallel, each on its own copy of the repo, and help you review and merge their work.

Pick by situation

Tap the ones that are you — the tools that fit light up below.

You want to see and steer each change inside your editorCursor Devin Desktop (formerly Windsurf)
You're comfortable in a terminal and want to hand off whole tasksClaude Code OpenAI Codex
You want to switch between model providers freely or keep your tools open sourceOpenCode
You don't want to change editorsGitHub Copilot

The contenders

Grouped by the side of the choice they answer, not ranked. Open a row for when to use it, the trade-off and what makers say.

ToolBest for

AI-first editor2

Cursorapp · free tierA VS Code-style editor with AI completions, chat and agents built in, keeping your VS Code extensions.

538$20Pro

Use it whenYou like working in an editor and want to see and steer each change as it happens.

Trade-offHeavy agent use is billed through credits, so costs can grow beyond the base plan.

llms.txt

Used by AdCreative.ai, Agentplace.io, AI Context Flow, ai_licia, Airbook AI and 533 more.

Cursor in full →Plans and calculator →vs Devin Desktop (formerly Windsurf) →vs Claude Code →

Cursor was our main AI assisted IDE for coding. Writing complex backend logic became so much faster. It literally feels like having a senior developer building the product with you.

Askiva AI, the makerSep 2026 ↗

Devin Desktop (formerly Windsurf)app · free tierA VS Code-style editor with unlimited Tab completions, an in-editor agent, and one place to manage local and cloud Devin agents.

43$20Pro

Use it whenYou want an AI editor with a free tier and plan to hand some tasks to Devin cloud agents.

Trade-offRecently renamed from Windsurf to Devin Desktop and reshaped around Cognition's agents; paid plans run on quotas, with extra usage at API rates.

llms.txt

Used by Aritect, Atlas, ComputerX, CreatorCube AI, Finyuus and 38 more.

Devin Desktop (formerly Windsurf) in full →Plans and calculator →vs Cursor →vs Claude Code →

Windsurf is amazing. It's our go to code editor and is amazing to use to help us build out new features and fix bugs in our application! Its especially amazing to use with Claude!

CreatorCube AI, the makerSep 2026 ↗

Terminal agent3

Claude CodeappHanding off multi-step tasks — features, refactors, bug hunts — to an agent that reads the repo, edits files and runs commands.

737+88 open source$20Pro18.9M/wk

Use it whenYou're comfortable in a terminal and want the agent to do the work while you review.

Trade-offUses Anthropic models only, and needs a Claude subscription or API credits.

llms.txt

Used by Accordio AI, Adventory, Agent 37 Cloud, agent-manager, Agent-Sin and 732 more · in 88 open-source projects.

Claude Code in full →Plans and calculator →vs Cursor →vs Devin Desktop (formerly Windsurf) →

We build HarnessRouter with Claude Code, and HarnessRouter runs Claude Code as a product backend for other teams. It is both our daily driver and one of the harnesses we route to.

HarnessRouter, the makerSep 2026 ↗

OpenAI Codexapp · open source · free tierDelegating tasks to an agent in the terminal, IDE or desktop app, or to cloud tasks that run in the background and open pull requests.

136$8.00ChatGPT Go★ 126.3k

Use it whenYou already pay for ChatGPT, or want an open-source CLI backed by OpenAI models.

Trade-offBuilt around OpenAI models, and usage limits depend on your ChatGPT plan.

llms.txt

Used by agent-manager, Agent-Sin, AgentRail, Almanac, Amami and 131 more.

OpenAI Codex in full →Plans and calculator →vs Cursor →vs Devin Desktop (formerly Windsurf) →

Codex is one of the three harnesses HarnessRouter routes to, and we use it in our own build loop. Its structured output made it one of the cleanest harnesses to normalize against.

HarnessRouter, the makerSep 2026 ↗

OpenCodeapp · open source · free tierAn open-source coding agent that works with almost any model provider, local models, or an existing Copilot or ChatGPT subscription.

42$10Go★ 210.3k

Use it whenYou want to switch models freely or keep your agent tooling open source.

Trade-offResult quality depends on the model you connect, and you manage provider keys and costs yourself.

CLI

Used by agent-manager, Blop AI, Brev.io, Cosyra, Cushion and 37 more.

OpenCode in full →Plans and calculator →vs Cursor →vs Claude Code →

A strong open local coding-agent runtime. We like it because it keeps agent work close to the developer environment while still fitting into a shared workspace through Vokal.

Vokal, the makerSep 2026 ↗

Editor plugin1

GitHub Copilotapp · free tierAdding AI completions, chat and agents to the editor you already use, including VS Code and JetBrains IDEs, plus pull request reviews on GitHub.

26$10Pro

Use it whenYou don't want to change editors, or you want to start on a free tier.

Trade-offDeeper agent features vary by editor, and heavy use counts against plan limits.

llms.txt

Used by Beatoven.ai, Billy.sh, Bolt, Edgee, Kastra and 21 more.

GitHub Copilot in full →Plans and calculator →vs Cursor →vs Devin Desktop (formerly Windsurf) →

Our token compression technology has to be compatible and efficient with Github Copilot, so we have to use it extensively to understand how we can optimize its context.

Edgee, the makerSep 2026 ↗

Cost: the cheapest plan that fits developers 1 seat, from list prices. Try your own numbers →

Cost as you grow

Each contender's cheapest usable plan as usage rises.

Who switches to what

Public pull requests on GitHub since Oct 2024 whose title says "X to Y" — real code changes, by developers in general rather than makers only. Pick a flow to see its pull requests.

Cursor →

Claude Code49 PRs · Compare →

[Common] Migrating the Cursor rules and commands to Claude Codea1exevs/set-forge · 2026-09-29
Move agent guidance from Cursor rules to Claude Codeadam-tabaczynski/eggs-delivery-system · 2026-09-26
Port .cursor skills and branch-safety rule to Claude Codeninadrathod/local-studio · 2026-09-22
Migrate PR review guidelines from Cursor BugBot to Claude Code Reviewduckduckgo/apple-browsers · 2026-09-21
chore/ACCT-0/ACCT-0/Port Cursor rules and skills to Claude Code config.omids0/accounting-with-google-sheet · 2026-09-19
Fill in CLAUDE.md and migrate Cursor framework to Claude CodeReforMan/Personal-Professional-Website · 2026-09-15
fix(ci): allow the cursor bot to trigger Claude Code ReviewWebFirstLanguage/wfl · 2026-08-31
Allow Cursor bot to trigger Claude Code Reviewyjsoon/howmuch · 2026-08-22
Vendor nine pstack skills, ported from Cursor to Claude Codegaryyau/agent-setup · 2026-08-20
Port Cursor pstack and thermos agent stack to Claude Code as kaizen-modepragmaticivan/dotfiles · 2026-08-17

All matching pull requests on GitHub ↗

Before you choose

How to approach it

Keep a short instructions file in the repo (stack, conventions, commands) so the assistant starts with context. Give it small, well-scoped tasks, have it run tests or the app, and read every diff before you commit. Try one tool properly for a week before switching; most of the gain comes from learning how to prompt and review.

Common mistakes

Accepting large diffs without reading them, so bugs and security holes you never saw end up in your codebase.
Letting an agent call a task done without running the tests or the app, so plausible-looking code that doesn't work lands in main.
Giving an agent production credentials or unrestricted shell access on your main machine, where one wrong command can do real damage.