Innovation & Tech

Apify Nears $100M ARR as AI Agents Fuel Tool Demand

Apify is approaching $100M gross ARR after AI agents boosted demand for its 90,000 tools. The platform paid creators about $2M last month, CEO Jan Čurn said.

By Amara Osei

4 min read

Jan Čurn, Co-Founder and CEO of Apify – Interview Series
Jan Čurn, Co-Founder and CEO of Apify – Interview Seriesstyles66 / pixabay

What's News

  • Apify is quickly approaching $100M gross ARR after agent usage took off in late 2025.
  • Apify paid a few hundred creators about $2M in total net earnings last month.
  • The platform hosts more than 90,000 tools, and over 300 community Actors earn over $1,000 per month.
  • Apify adopted Anthropic's Model Context Protocol at the end of 2024.
  • The company was founded by Jan Čurn and Jakub Balada in 2015 through the Y Combinator Fellowship.

Apify's usage has taken off since late 2025 and the platform is quickly approaching $100M in gross ARR, Co-Founder and CEO Jan Čurn said. Last month alone, Apify paid a few hundred creators about $2M in total net earnings.

Čurn launched the company with Jakub Balada in 2015 through the Y Combinator Fellowship. The original product solved a problem the two founders hit themselves: a client project needed listings of new apartments in Prague, and existing tools struggled with dynamic JavaScript and irregular website structures. They built a tool to turn any website into an API.

A decade later, Apify spans web scraping, browser automation, and a marketplace of ready-to-run tools called Actors. More than 90,000 tools now cover web data, scraping, automation, and AI. A growing share of the people using them aren't people at all — they're AI agents.

When did scraping become an agent business?

The first shift came in early 2023, soon after ChatGPT launched. "Suddenly you could feed fresh web data into a prompt instead of training a new model," Čurn said. Apify was among the first scraping platforms to provide crawlers for emerging RAG pipelines and integrate with libraries like LangChain and LlamaIndex.

Then came the agent wave. Ask an agent to compare a product's price across 20 e-commerce sites, and where does it get the data? That was the problem Apify had been solving for developers for years. When Anthropic released the Model Context Protocol (MCP) at the end of 2024, Apify adopted it right away. "We already had thousands of tools with a clear input and output, which is exactly what an agent needs," Čurn said.

The real turning point came at the end of 2025, when models got good enough for agents to do meaningful work.

Why would an agent rent instead of build?

Čurn argues the models will keep generating ever more complex software from scratch, but there's a threshold above which it's more efficient for an agent to buy an existing tool rather than build it. That threshold will move up as models improve, but some will always remain — leaving room to sell pre-built tools, especially those based on private data and hard-to-get context.

"Rather than paying $500 to generate it properly, you can pay someone $5 to rent an existing service," he said. As the cost of writing software falls and tool counts grow, the main constraint is becoming discovery: finding trusted, valuable tools for the job.

Production use cases on the platform are oddly specific. One AI setup scans YouTube, eBay, and Japanese shops to decide which Pokémon cards are worth buying. An insurer used Apify to check 55 Italian obituary sites to verify policyholder deaths. Canary uses Apify for Spotlight, a tool that helps investigators find victims of child trafficking.

What breaks when agents hit the open web?

An LLM, in Čurn's framing, is a lossy compression of its training data. "It can make very clever guesses, but it's still guessing," he said. Prices change, products disappear, and websites update constantly, so decision-making agents need a reliable way to get current web information as structured data.

The technical picture is unforgiving. Modern websites return dynamic JavaScript, challenge pages, or incomplete content, while anti-bot systems inspect everything from IP reputation to browser fingerprints. Failures are often silent: "an agent can confidently summarize an empty page without noticing," Čurn said. Building something that works once is easy; keeping it working at production scale is much harder.

Permissions grow critical once agents can act, not just read. The agent should never see user credentials, and all secrets must live outside the model's reach. Apify has been working on this proxy sandboxing problem in Actors and in mcpc, its open-source CLI client for MCP.

Where does the agent economy go from here?

Čurn sees platforms like ChatGPT and Claude becoming a new operating system, with tools plugging in like phone apps. Agents that need mapping data can either click through Google Maps in a browser — imprecise and expensive — or call a tool that returns structured data faster and more reliably. Main agents will also hand off tasks like booking tables or finding flights to specialized sub-agents on cheaper models.

He also sketches an agent-payment future: give an agent a task and a small wallet, and let it pay for the tools it needs using a protocol like x402 or MPP. The volume is still small, but Čurn believes most economic activity will eventually happen autonomously between agents, with ever lower human oversight.

On the marketplace side, over 300 community Actors now generate more than $1,000 per month for their creators. Čurn's stated goal for Apify's second decade: become the world's largest and most trusted marketplace of AI tools for people and their billions of future agents.

Source: unite.ai

Share this article:

More from Amara Osei

Amara Osei

Show full bio

Senior reporter covering companies, deals and strategy in entrepreneurship, startups and venture funding at Business Bearings.

2,336 articles

Related articles

« Previous articleNext article »