Exclusive: Browserbase launches headless browser platform that lets LLMs automate web tasks

Browserbase, a San Francisco startup, today announced the launch of its platform that enables developers to automate complex web tasks using AI and headless browser technology. The company aims to streamline the process of web automation, making it easier for developers to create and manage sophisticated workflows without the need for extensive infrastructure setup. Browserbase is opening up self-serve access to its platform today, allowing any developer to start automating web interactions.

The startup also announced it has raised $6.5 million in seed funding. The round was led by Kleiner Perkins, with participation from Basecase Capital, AI Grant, and angel investors including Vercel CEO Guillermo Rauch, Pagerduty cofounder Andrew Miklas, and Hex CEO Barry McCardel. 

Solving the headless browser pain point

Founded by Paul Klein, a seasoned entrepreneur and developer, Browserbase emerged from Klein’s recurring frustrations with headless browser infrastructure in his previous roles. Klein’s journey in the tech industry began at Twilio, before leaving to found Stream Club, a live-streaming startup that aimed to simplify the process of creating and sharing live content. In 2021, Klein sold Streamclub to Mux, a leading video API company, Despite his successful exit, Klein remained acutely aware of the pain points developers faced when working with browser automation, a challenge he frequently encountered at Streamclub.

“Headless browser infrastructure was always such a painful but necessary primitive,” says Klein. “Fast forward two years after my last acquisition, I started to see more and more AI companies needing headless browsers to automate web tasks and retrieve data. I had the expertise, and the product I needed still didn’t exist – so I set out to build it with Browserbase.”

Browserbase provides a programmable browser platform that allows developers to automate complex online tasks using code. It offers features like advanced debugging, session recording, a proxy supernetwork, and bot detection avoidance to streamline web automation for compatibility with popular tools like Puppeteer, Playwright, and Selenium. The startup aims to serve as a key building block in an emerging AI software stack.

“It’s not going to be just developers writing this code, it’s going to be large language models generating this code longer term,” Klein explains. “So it’s both developers and LLMs controlling these web browsers to go out and automate the daily tasks we do online.”

The AI automation opportunity

While headless browsers have often been used for web scraping, Klein sees Browserbase as solving a broader need around task automation that has become more tractable with advances in AI. Klein points out that Browserbase can be used for web scraping, but it’s best suited for automating workflows, stating: “It’s not just getting the contents of the page. You want to add this item to cart. You want to put in your zip code and see how long it’s going to take to ship.”

Further, Klein sees Browserbase as a key component of what Andrej Karpathy, former director of AI at Tesla, calls the “LLMOS” (Large Language Model Operating System). In this ecosystem, large language models serve as the central “brain,” while tools like Browserbase act as the “eyes and ears,” enabling AI to perceive and interact with the web.

“Andrej Karpathy has this idea called the LLOS,” Klein explains. “The LLM is like the CPU, the vector database is like the memory, and the browser is part of the sensory input, like the visual, video, audio, and text input. So I really think that we’re building this core part of the LLMOS.”

By positioning Browserbase as a critical component of the AI ecosystem, Klein believes the platform will play a crucial role in enabling the next generation of AI-powered applications and automations.He likens the current AI automation opportunity to Twilio’s growth on the back of two-factor authentication going mainstream:

“Right before Twilio’s IPO, everyone started adding two-factor auth and receiving text messages to sign-in. That was a huge driver for them,” he notes. “I would point to the rise of AI and large language models as our ‘2FA moment’ – this is our time to be loud because there’s going to be an explosion of AI apps built in the next year.”

As more businesses seek to harness the power of AI and automation, Klein believes that the demand for tools like Browserbase will continue to grow, positioning the company as a key player in this emerging market.

Enabling enterprise AI developers

Browserbase provides robust integrations with popular AI/ML tools and frameworks like Langchain, LlamaIndex, Crew AI, and Vercel. It offers SDKs in Python and Node.js. The platform can automate both public websites and internal authenticated enterprise applications.

Looking ahead, Klein believes Browserbase can ultimately serve as a “last-mile” solution to AI-enable the massive long-tail of the web. 

“There are so many websites that will never be updated – it’s like rings on a tree with the oldest websites at the center,” he says. “I don’t think we’ll build an API over the entire web for AI agents. I think AI will become smart enough to use the web on its own. And if we want to support the legacy web, we’ll always need browsers.”

Browserbase’s platform has significant implications for enterprise developers tasked with automating internal applications that often have inadequate or outdated APIs. Many enterprises rely on legacy systems that lack robust, modern APIs, making it challenging to integrate and automate workflows.

“In large enterprises, there are so many internal applications that will never be updated with modern APIs,” Klein explains. “Developers are still expected to automate these systems, but traditional tools fall short. Browserbase allows them to automate these antiquated applications as if they were interacting with them in a browser, bypassing the need for direct API access.”

Importantly, Browserbase offers an on-premise deployment option, allowing enterprises to run the platform behind their own firewalls. This is crucial for organizations with strict data governance and security requirements, as it ensures sensitive data remains within their controlled environment.

“We understand that many enterprises need to keep their data on-premise for compliance reasons,” Klein notes. “That’s why we offer a seamless on-premise deployment option. Enterprises can have their own private instance of Browserbase running behind their firewall, giving them full control and ownership of their data.”

By combining the power of headless browser automation with the security of on-premise deployment, Browserbase enables enterprise developers to streamline complex internal workflows while adhering to strict data governance policies. This positions Browserbase as a compelling solution for enterprises looking to modernize their operations without compromising on security.

Pricing and Availability 

As part of today’s announcement, Browserbase is opening up self-serve access to its platform, allowing developers to start automating web interactions immediately. The company offers three pricing tiers to accommodate different needs and scales:

1. Hobby Plan ($39/month): Includes 200 browser hours, 2 GB of proxies, 3 concurrent browsers, 7-day data retention, auto captcha solving, and basic stealth mode.
   
2. Startup Plan ($99/month): Includes 500 browser hours, 5 GB of proxies, 50 concurrent browsers, 30-day data retention, auto captcha solving, and advanced stealth mode.
   
3. Enterprise Plan (Custom Pricing): Includes self-hosting options, dedicated proxies, 100+ concurrent browsers, 90+ day data retention, auto captcha solving, and advanced stealth mode.

“We believe in providing a high-quality, reliable service that developers can trust,” Klein emphasizes. “While we don’t offer a free tier, our pricing plans are designed to provide value for developers and businesses of all sizes. We encourage anyone interested in automating web interactions to sign up and start building with Browserbase today.”

Developers can sign up for Browserbase and start automating web interactions by visiting the company’s website at www.browserbase.com.

Scaling up with fresh funding 

With $6.5 million in new funding, Browserbase plans to expand its engineering team while remaining focused on product development. The new funding will also be used to accelerate product development and enhance market reach as the company aims to become a cornerstone of AI applications.

One unique aspect of Browserbase’s approach is the company’s commitment to building an all in-person team. Klein believes that having everyone work together in the same physical space fosters better communication, collaboration, and culture.

“We’re based in San Francisco and I’m going to just show you that really quick,” Klein says as he turns his camera to show the whole team working together in their small office. “We’re all in person.” Everyone waves to me. 

Klein acknowledges that this approach has made hiring more challenging, as it requires candidates to be willing to commit to working in-person for 40-60 hours per week. However, he believes that this has become a key part of Browserbase’s culture and has helped them attract people who are truly bought into their vision and way of working.

“Hiring in-person creates a higher bar – candidates have to be willing to commit to being here 40-60 hours per week,” Klein notes. “It made hiring more challenging at first, but it’s become part of our culture. We’re pulling in people who are opting in to being in-person.”

As Browserbase opens its doors to developers everywhere, Klein is excited to see what people build but cognizant they are just one part of an emerging AI stack. Still, he views the launch as validation for an under-appreciated infrastructure problem he long struggled with as a developer.

“Headless browsers felt like my own little world, this one problem I was obsessed with,” Klein reflects. “Having customers now say ‘This has been very helpful for us’ – that’s been really validating. There are dozens of us out there who’ve felt this pain. It’s cool to meet them and hopefully save them some headaches with Browserbase.”