Claude Code

Anthropic's agentic coding CLI that runs in a terminal, reads and edits a codebase, runs commands, and completes multi-step engineering tasks.

Works with AI agents:llms.txt
Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗

Claude Code is Anthropic's agentic coding tool. You describe a task in plain language, and it reads the codebase, edits files across the project, runs commands such as tests and git, and reports back. The CLI runs on your machine and talks directly to the model API, with no Anthropic backend server or remote index of your code in between.

It comes as a terminal CLI for macOS, Linux and Windows (native or WSL), an extension for VS Code and Cursor, a JetBrains plugin, a desktop app, and a web version at claude.ai/code where sessions run in the cloud. In manual mode it asks before editing files or running commands; auto mode lets a classifier approve safe actions instead, and an optional sandbox isolates shell commands from the filesystem and network. Project instructions live in CLAUDE.md (it can also read AGENTS.md), and you extend it with MCP servers, skills, hooks and plugins.

You sign in with a Claude subscription or an Anthropic Console API key. The CLI, VS Code and JetBrains integrations also run against Amazon Bedrock, Google Cloud or Microsoft Foundry, which lets a company bill and keep inference in its own cloud account. Web sessions handle long tasks in parallel, including on repositories you don't have locally.

It needs an internet connection and works only in Anthropic's supported countries. Command sandboxing needs WSL 2 rather than native Windows, and on consumer plans your sessions can be used for training unless you turn that setting off.

Where it fits

Who uses it

738 makers' products, each linked to the source that shows it, and 88 open-source projects that declare it in their code.

The maker says so 738Declared in code 90How evidence is collected →
TwentyOpen-source CRM for managing contacts, companies, and deals, built as a self-hostable al…

“Thanks for helping us build the website that fast!”

AI Coding AssistantsMaker says so +1 · source ↗
giselleGiselle is an open-source tool for building and running AI workflows in a visual node ed…

“Claude Code has transformed how we build Giselle. It understands our monorepo architecture, coding conventions, and domain concepts — making implementation feel like pair programming with a senior engineer who knows the entire codebase. Code reviews that used to take hours now take minutes as Claude spots edge cases and suggests improvements. Our development velocity has never been higher. Thank you for building the future of software development!”

AI Coding AssistantsMaker says so · source ↗
GrowthBookOpen-source feature flagging and A/B testing platform with built-in statistical analysis…

“Claude Code helps our engineering team move faster inside the actual repo. It is especially useful for debugging, refactoring, writing tests, and getting oriented in complex code without constantly switching contexts.”

AI Coding AssistantsMaker says so · source ↗
supersetDesktop IDE for running several coding agents in parallel, each in an isolated workspace…

“We use Claude Code to build Superset. It is also one of the terminal agents our users can orchestrate from Superset on desktop and iPhone.”

AI Coding AssistantsMaker says so · source ↗
BrandingStudio.ai®Agency-quality branding in 60 minutes, not 6 months

“Claude Code with Opus is the backbone of my development workflow. The depth of context it provides across a complex, multi-module platform like BrandingStudio.ai is genuinely unmatched; it has significantly accelerated my build and continues to power my iteration speed.”

AI Coding AssistantsMaker says so · source ↗
MindPalMindPal is a no-code platform for turning a person's expertise into AI agents and multi-…

“Claude Code is how we build MindPal itself.”

AI Coding AssistantsMaker says so · source ↗
KatalystThe AI agent that works your Salesforce Pipeline

“We tried Cursor and moved to Claude Code, it's just fit how we build better. The skills piece is the difference, we can teach it our own patterns and it actually holds them across the codebase. Tip - Invest early in setting up your own skills instead of using it out of the box. The more you teach it how your codebase actually works, the more it feels like a teammate who knows your patterns rather than a generic autocomplete.”

AI Coding AssistantsMaker says so · source ↗
CekuraAutomated QA for Voice AI and Chat AI agents

“Skills to design and optimize metrics easily - Can refer to skills here: https://github.com/cekura-ai/cla...”

AI Coding AssistantsMaker says so · source ↗
DataboxAgentic analytics platform for you and your agents.

“MCP Connectors means implementing an open protocol correctly: parsing tool schemas from any server a user points us to, handling three different auth flows (OAuth, API key, bearer token), and building permission logic that fails safe by default, since this feature can give an AI real access to a CRM, a support desk, or an inbox. That's well-scoped, spec-driven engineering work. We built the connector framework with Claude Code, working through the MCP spec and the edge cases (malformed schemas, expired tokens, servers that don't return what they promise) fast enough to ship a working connector framework, official connectors, and a custom connector option in the same release.”

AI Coding AssistantsMaker says so · source ↗
Klariqo AI Voice AssistantsAI Voice assistant in 3 minutes. Built for non-developers

“Internally, some team members refer to Claude code as our virtual CTO. Code improvements, suggestions, and quick bulk fixes are all handled by Claude code.”

AI Coding AssistantsMaker says so · source ↗
Perfai SecurityFind & fix live vulnerabilities in Vibe Apps with 1-prompt.

“We compliment Claude Code with our security solution built for vibe-coders using the platform.”

AI Coding AssistantsMaker says so · source ↗
PopAirYour All-in-one AI Copilot for MacOS

“Superior code generation and reasoning capabilities accelerated our development significantly—Claude Code became an essential part of our workflow.”

AI Coding AssistantsMaker says so · source ↗
serenitiesConnect your own AI with unlimited prompts and easy deploy

“Best reasoning for complex codebases. Opus handles multi-file refactors that others can't.”

AI Coding AssistantsMaker says so · source ↗
ClientJamAI-powered lead generation for designers and agencies

“ClientJam was built solo — with Claude Code as my pair-programmer the whole way. It wasn't "AI wrote my app" so much as a constant collaborator that let one person move like a small team”

AI Coding AssistantsMaker says so · source ↗
EdgeeOne gateway for cheaper, faster, unstoppable coding agents

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

AI Coding AssistantsMaker says so · source ↗

Open source: a project that declares Claude Code as a dependency in its public code — verifiable, but not necessarily a live product.

What makers say

533 makers on why they use Claude Code, in their own words on Product Hunt.

I've used Cursor, Copilot, Windsurf, and Bolt — they're all solid for code completion and file-level edits. But Claude Code is the only tool where I could say "build me a SaaS app" and actually get a production-ready application. DwellRecord went from an empty folder to a live, paying product at dwellrecord.com in a single week. Not a prototype — a full app with Google OAuth, Stripe subscriptions, PostgreSQL, PDF exports, AI-powered receipt scanning, email verification, an admin portal, and security hardening. All deployed to Railway with Cloudflare in front. The difference is Claude Code doesn't just write code — it architects, debugs, deploys, and iterates. When our CSP headers broke analytics scripts, it diagnosed the console errors and fixed them. When Cloudflare's proxy caused CSRF mismatches on form submissions, it traced the 403s to a www/non-www origin mismatch and resolved it. It thinks like a senior engineer, not an autocomplete. The other tools help you code faster. Claude Code helps you ship.
DwellRecordSep 2026 ↗
I've used Cursor, Copilot, Windsurf, and Bolt — they're all solid for code completion and file-level edits. But Claude Code is the only tool where I could say "build me a SaaS app" and actually get a production-ready application. DwellRecord went from an empty folder to a live, paying product at dwellrecord.com in a single week. Not a prototype — a full app with Google OAuth, Stripe subscriptions, PostgreSQL, PDF exports, AI-powered receipt scanning, email verification, an admin portal, and security hardening. All deployed to Railway with Cloudflare in front. The difference is Claude Code doesn't just write code — it architects, debugs, deploys, and iterates. When our CSP headers broke analytics scripts, it diagnosed the console errors and fixed them. When Cloudflare's proxy caused CSRF mismatches on form submissions, it traced the 403s to a www/non-www origin mismatch and resolved it. It thinks like a senior engineer, not an autocomplete. The other tools help you code faster. Claude Code helps you ship.
DwellRecordMar 2026 ↗
MCP Connectors means implementing an open protocol correctly: parsing tool schemas from any server a user points us to, handling three different auth flows (OAuth, API key, bearer token), and building permission logic that fails safe by default, since this feature can give an AI real access to a CRM, a support desk, or an inbox. That's well-scoped, spec-driven engineering work. We built the connector framework with Claude Code, working through the MCP spec and the edge cases (malformed schemas, expired tokens, servers that don't return what they promise) fast enough to ship a working connector framework, official connectors, and a custom connector option in the same release.
DataboxSep 2026 ↗
Claude Code was the backbone of building OpenCut AI. We added 8+ AI features transcription, voice cloning, filler removal, AI image generation, natural language editing on top of an existing open-source video editor, and Claude Code's deep codebase understanding made it possible. Unlike Cursor or Copilot, it doesn't just autocomplete lines it reasons across the entire project, refactors confidently, and handles multi-file changes spanning a Next.js frontend and FastAPI AI backend in one shot. The 1M context window meant we could feed it our whole architecture and get coherent, production-ready code back. Nothing else came close for a project this ambitious
Free AI Video Editor OpenCutAISep 2026 ↗
Claude Code stood out because it handled complex architecture, product design, and technical planning more professionally than the alternatives. I was working on a project that involved web, desktop, and macOS compatibility. When I explained complex requirements, Claude Code understood the bigger picture, broke the work into clear steps, and helped me plan the system in a complete way. What mattered most was that it did not only generate code. It helped me organize the visual structure, product flow, architecture, and implementation plan. When I shared messy ideas or complicated workflows, it turned them into a clear and professional development roadmap.
Termi ProtocolSep 2026 ↗

Loved and watch-outs

Themes that recur in makers' words and Hacker News comments, each linked to what it summarises, with how Product Hunt tags its reviews.

Most loved
  • Holds deep context across large multi-module codebases, so one change can span UI, backend, docs and infrastructure. PHPH 2PH 3PH 4
  • Carries complex multi-step tasks through end to end, such as waiting on CI, reading logs and iterating until green. PHHNHN 2PH 2
  • Hooks, skills, subagents, plugins and MCP make it a host that teams extend and build products on. PHPH 2HNPH 3
  • Subscription plans give much more usage than the same money spent on API tokens. HNHN 2HN 3
Watch-outs
  • Quality can shift between releases because of harness changes, such as a quietly lowered default reasoning effort. HNHN 2HN 3HN 4
  • Subagents are opaque and hard to steer, and the default setup feels heavy, so power users strip tools and prompts. HNHN 2HN 3HN 4
  • The subscription only works inside Claude Code, not in other harnesses people prefer. HNHN 2HN 3
  • The CLI is resource hungry, using hundreds of megabytes of RAM per idle session. HNHN 2
On Product Hunt 5.0★ · 736 reviews
code generation 11complex software tasks 8massive context window 2
Read the reviews on Product Hunt ↗

Reliability and open issues

Incidents on its status page50+ in the last 90 days, 21 major · 50+ in the last yearFrom its own status page, 2026-10-02. Vendors decide what they post; one page can cover several products.

Who switches

Public pull requests on GitHub since Oct 2024 whose title says "X to Y" — real code changes moving a project from one tool to another, by developers in general. Open a row to see the pull requests.

Cursor → Claude Code49 PRs
OpenAI Codex → Claude Code45+ PRs
GitHub Copilot → Claude Code40 PRs
OpenCode → Claude Code28 PRs
Claude Code → OpenCode44 PRs
Claude Code → OpenAI Codex34+ PRs
Claude Code → GitHub Copilot28 PRs
Claude Code → Cursor10 PRs

Alternatives to Claude Code

All alternatives by situation →

On Product Hunt, people weigh it against: Cursor, OpenAI Codex, OpenCode, Kilo Code, Google Antigravity, Augment Code, Devin Desktop (formerly Windsurf).

Questions makers ask about Claude Code

Does it work in VS Code or JetBrains IDEs?

Yes. There is an extension for VS Code (also installable in Cursor) and a plugin for IntelliJ IDEA, PyCharm, WebStorm and other JetBrains IDEs; the JetBrains plugin needs the Claude Code CLI installed separately. source ↗

Is my code used to train Anthropic's models?

On Free, Pro and Max plans it depends on your privacy setting, and data is used for training only when that setting is on. Under commercial terms (Team, Enterprise, API and cloud platforms) Anthropic does not train on your code or prompts unless your organization opts in. source ↗

Where does my code go?

Claude Code runs locally and sends your prompts and the model's outputs to the model API over TLS. There is no separate Anthropic server holding an index of your repository. source ↗

Can we run it through AWS, Google Cloud or Azure?

Yes. It supports Amazon Bedrock, Google Cloud's Agent Platform and Microsoft Foundry, using their regions, credentials and billing, and you can route traffic through a corporate proxy or LLM gateway. source ↗

Does it run on Windows?

Yes, natively from PowerShell or CMD, or inside WSL. Command sandboxing is only supported under WSL 2, and Git for Windows is recommended for native installs so it can use Bash. source ↗

Can it work offline?

No. It requires an internet connection and is available only in Anthropic's supported countries. source ↗

Does it read AGENTS.md?

Yes. Alongside CLAUDE.md files, Claude Code can load AGENTS.md, so one instruction file can be shared with other coding tools. source ↗

Is Claude Code free?

No permanent free tier; paid use starts at $20/mo. source ↗

Is Claude Code open source or self-hostable?

Not open source, and hosted only.

Can AI coding agents work with Claude Code?

It serves an llms.txt docs index.

Who uses Claude Code?

738 makers' products we track, each with a source, and 88 open-source projects declare it in their code. source ↗