The Million Club — Every AI Coding Tool, IDE and Agent Ranked by Real Traffic

The Full Rankings

Here are all 76 AI coding, IDE, and agent tools ranked by monthly traffic. Every single one offers a free tier. The scale ranges from GitHub Copilot's 504 million down to Rork's 781 thousand — and every tool in between represents a real, active community of developers choosing AI-assisted development.

The IDE Revolution

GitHub Copilot at 504 million monthly visits sits so far above everything else on this list that it barely belongs in the same ranking. Half a billion visits. That's more than the next fifteen tools combined. Copilot didn't just win the AI code assistant market — it defined it, embedded itself in the world's most popular code editor, and became the baseline that every competitor measures against.

But the real story in 2025 isn't Copilot's dominance — it's the AI-native IDE explosion that's challenging the entire VS Code ecosystem from the ground up.

The absent giants deserve mention. Claude Code from Anthropic has become a developer favorite for its deep reasoning and agentic coding capabilities, but it runs through the Claude interface and CLI without its own domain. OpenAI's Codex agent and Google's Gemini CLI are in the same position — massively used, invisible to web traffic metrics. If these tools had standalone domains, the top ten of this ranking would look very different.

The App Builder Explosion

The most dramatic traffic story in this entire ranking belongs to Lovable. At a combined 45.39 million visits across two domains, it appeared seemingly out of nowhere to become the most-visited AI app builder on the internet. Describe what you want — "a project management tool with kanban boards and team chat" — and Lovable generates the full-stack application, deploys it, and hands you the code.

This "describe and deploy" model has exploded into an entire category. Bolt at 6.44 million combined, Replit at 12.7 million combined, Vercel's v0 at 5.3 million, Emergent at 4 million, and WebSim at 1.47 million — each takes a slightly different angle on the same premise. Bolt emphasizes speed and simplicity. Replit wraps generation in a full development environment with hosting. v0 specializes in UI components and frontend code. The common thread: you describe, AI builds.

The website builder incumbents are adapting fast. Wix at 34.22 million has layered AI generation throughout its platform — from page layout to copy to image generation. Webflow at 6.4 million serves the design-focused crowd with AI-assisted responsive layouts. Framer at 4.75 million targets the "design-to-site" workflow with AI that understands both visual design and code generation.

The no-code platforms tell a different story. Airtable at 22 million, PowerApps at 14.1 million, Bubble at 3.74 million combined, and Glide at 2.82 million combined have been letting non-developers build applications for years. AI supercharges their proposition — what used to require careful drag-and-drop configuration now happens from a text description. Retool at 1.66 million and Knack at 1.21 million serve enterprise internal tools with the same AI-augmented approach.

The Agent Uprising

Manus at 16.64 million combined visits is the breakout agent story of 2025. While other AI agents remained demos and research papers, Manus shipped a product that actually works for non-trivial tasks — research, data analysis, web browsing, code execution, all orchestrated by an AI that plans and executes multi-step workflows autonomously.

The workflow automation category is its own mature ecosystem. n8n at 15.22 million combined has become the open-source darling of the AI automation world. Its visual workflow builder, combined with deep AI model integrations, lets developers and power users build sophisticated automations without writing much code. Zapier at 5.9 million — the original workflow automation tool — has layered AI throughout its platform. Make at 5.17 million competes directly with deeper visual automation capabilities. UiPath at 1.73 million brings AI to enterprise robotic process automation.

What separates the 2025 agent landscape from earlier automation tools is autonomy. Traditional automation runs predefined workflows. AI agents decide their own approach, adapt when things go wrong, and handle ambiguity. Manus can research a topic, synthesize information from multiple sources, write a report, and generate visualizations — without someone specifying each step. That's a qualitative leap, not just an incremental improvement.

Pipefy at 2.15 million and Dify at 1.19 million represent the "build your own agent" approach — platforms where you can construct custom AI agents tailored to specific business processes. LangChain at 2.94 million provides the developer framework that many of these tools are built on — the React or Rails of the AI agent world.

The ML and AI Ecosystem

Hugging Face at 24.03 million is the GitHub of machine learning — the place where models, datasets, and research converge. Every major AI model eventually lands on Hugging Face, and the platform has become the default infrastructure for anyone doing serious ML work. Its traffic reflects not just popularity but necessity: if you're building with AI, you're almost certainly using Hugging Face at some point.

Kaggle at 11.38 million serves the learning and competition side of AI. Its datasets are used by researchers and students worldwide, and its competitions have launched thousands of ML careers. Google Colab at 2.71 million provides the free GPU-powered notebooks that make ML accessible to anyone with a browser — you can train a neural network without owning a GPU, and that accessibility has been transformative for ML education.

The local LLM movement deserves its own spotlight. Ollama at 4.73 million and LM Studio at 1.24 million serve developers who want to run large language models on their own hardware — no API keys, no usage fees, no data leaving their machine. The appeal is both practical (cost, latency, privacy) and philosophical (independence from cloud providers). The combined traffic of nearly 6 million shows this isn't a niche movement.

ModelScope at 3.25 million is Alibaba's answer to Hugging Face, serving the Chinese AI development community. Weights & Biases at 2.08 million provides experiment tracking — the "lab notebook" that makes ML research reproducible. Anaconda at 2.07 million, Scikit-learn at 1.18 million, and Streamlit at 1.33 million form the Python data science stack that predates the LLM era but remains foundational. Roboflow at 1.4 million and Ultralytics at 1.21 million own the computer vision niche, powering everything from factory quality control to autonomous vehicle perception.

The Infrastructure Layer

Every AI application needs somewhere to run, and the infrastructure layer of this ranking reveals who's powering the AI development boom behind the scenes.

The OpenAI Platform at 17 million is where developers build on GPT — the API documentation, the playground, the fine-tuning tools. Vercel at 15.23 million has become the default deployment platform for AI-powered web applications, partly through its own v0 code generation tool. GitLab at 17.16 million combined competes with GitHub on the DevSecOps front, with its Duo AI assistant integrated throughout the pipeline.

The GPU cloud market is white-hot. RunPod at 1.93 million and Groq at 1.94 million serve developers who need GPU compute for training and inference. Groq's custom LPU chips offer inference speeds that traditional GPUs can't match — and the 1.94 million visits suggest developers are noticing. SiliconFlow at 1.47 million and RunningHub at 1.13 million serve the Chinese market. Replicate at 1.66 million makes it trivially easy to deploy and run open-source models in the cloud.

DigitalOcean at 12.54 million combined remains the approachable cloud for developers who find AWS overwhelming. Heroku at 1.51 million — once the darling of startup deployment — has been revitalized with AI-aware features. Apify at 2.25 million occupies an interesting niche: AI-powered web scraping and data extraction, the unglamorous infrastructure that feeds training data to every other tool on this list.

Graphite at 1 million represents a new wave — AI-powered code review that catches bugs, suggests improvements, and accelerates the pull request process. Builder.io at 1.05 million uses AI to generate frontend code from designs. Warp at 1.13 million reimagines the terminal itself as an AI-native interface. These tools don't generate whole applications, but they make every step of the development process faster.

How to Choose Your Dev Tool

Every tool on this list offers a free tier — all 76 of them. The challenge isn't cost. It's figuring out which tool matches how you think about building software.

The honest advice: most developers end up using three or four tools from this list simultaneously. An AI IDE for writing code, an app builder for prototyping, a deployment platform, and some automation tool for repetitive workflows. The tools that win long-term will be the ones that make it easiest to move between these stages without switching context.

Methodology and Data Source

All traffic numbers come from SimilarWeb, reflecting December 2025 estimates.

This is the broadest ranking in the Million Club series. I've included AI code editors, coding assistants, app builders, website builders, no-code platforms, workflow automation tools, AI agents, ML platforms, cloud infrastructure, and developer tools. The common thread: every tool on this list uses AI to make some aspect of building software faster, easier, or more accessible.

Several important tools are absent due to the web traffic methodology. Claude Code (Anthropic), OpenAI Codex, and Gemini CLI are all popular AI coding tools that operate through parent platforms or command-line interfaces without standalone website domains. GitHub Copilot is tracked via its github.com/copilot path, but tools that live inside terminal sessions or larger platforms can't be measured by SimilarWeb. Their absence from this ranking doesn't diminish their significance — Claude Code and Codex are arguably among the most important AI coding tools in existence.

Some entries include multiple domains for the same product (Lovable, Cursor, n8n, Bolt, Replit, GitLab, Bubble, Glide, Manus, Trae, Rentry, DigitalOcean). I've listed each domain separately to maintain consistency with SimilarWeb's domain-level tracking. Where I reference combined traffic in the narrative sections, I note it explicitly.

Every tool on this list — all 76 — offers a free tier. This 100% free-tier rate matches the audio ranking and reflects the developer tool market's universal truth: developers try before they buy, and any tool without a free tier is invisible to them.