Bytebot simplifies and streamlines web scraping & automation by translating natural language prompts into intricate browser tasks, from data extracting to form filling. Get all the AI-enabled capabilities you need for your web workflow, without the complexity.
Thanks for the hunt @mwseibel !
Hey everyone, I’m Carl, an ex-Google and ex-Apple software engineer obsessed with building developer tools. After realizing how frustrating it is to build a browser scraping and automation app, I decided to build my own LLM-based toolchain to simplify the process. I’m excited to finally launch the result—Bytebot!
Bytebot is building the state-of-the-art web scraping toolchain, you can think of it as “Langchain for scraping” or “AI powered BeautifulSoup” written in TypeScript. It handles 80% of the work involved in building a web parser, so developers spend less time tediously parsing HTML and more time building their end application.
Bytebot offers the following features:
- 🦾 Automatic HTML semantic parsing into content and possible user actions
- 💼 Robust scraping scripts that are portable across vastly different websites
- ❤️🩹 Self-healing scripts that adapt to website layout changes or updates
From dynamic form filling, to automated data-extraction, to intelligent web interactions and more, Bytebot automates your time-consuming tasks. We're excited to equip other builders with our tools and see what they create!
To get started, sign up here: https://bytebot.ai/get-s....
To stay posted with the latest updates, follow Bytebot on LinkedIn: https://www.linkedin.com/company...
Congrats on the launch! Really curious to try it out, I've see many so called AI scrapers, and most of them failed at basic tasks.
Very curious about the pricing, if I read correctly, the limits are on the number of domains, not the number of pages/records extracted from each website?
Also... any plan for a cloud version that helps bypass captchas and various security systems?
Thanks for the support @massimocw, great questions!
Yeah, a lot of the end-to-end AI scrapers (which try to accomplish everything from a single prompt) end up struggling. We believe a more granular approach is the right abstraction.
Re: pricing, there is a basic cap on the # of requests on the individual plan. We can make that more clear!
A cloud version with advanced features is in the works! Happy to chat more, here's a link to schedule some time: https://calendly.com/bytebot/byt...
Hi Carl,
Congratulations on launching Bytebot! It's fascinating to see AI streamline such a complex process. The self-healing scripts feature sounds particularly impressive. Can't wait to see how it adapts to different web layouts. What was the biggest challenge you faced while integrating AI with browser automation?
Cheers for your hard work!
Thanks for the support @alex_dulub1!
Great question - the biggest challenge is definitely handling the myriad of ways information is represented on the web. There's like 10 different ways to make a table in HTML 😂
The benefit for Bytebot is that it improves with every new domain that it's exposed to! Happy to chat more, here's a link to find a time: https://calendly.com/bytebot/byt...
Report
A long time ago I was in a team that spent months trying to solve what this will now do in minutes. Nice work. Granular abstractions are the way to go @atupem !
i’ve tried countless automation apps for years including writing my own puppeteer scripted. this is the perfect middle ground of automation and control
Thanks for the support @sidhantgandhi! Exactly, we’re looking to provide developers with the right level of flexibility!
Report
This is a great solution to an obvious problem! Really like how you've simplified the code problem here. Very well done, looking forward to having this implemented! Got a few projects I'm thinking about in particular 🧠
Thanks for the support @trent_kennelly!
Glad to hear it resonates, I felt this pain myself on a number of occasions! What projects did you have in mind? Happy to chat more, here's a link to find a time: https://calendly.com/bytebot/byt...
Report
I can't tell you how many times I've had selenium scripts break on me for not apparent reason, or the nuance of trying to extract some data when the data is loaded in asynchronously. I'm goin to try this out asap.
Bytebot
Breadcrumbs
Bytebot
Breadcrumbs
Web3 Antivirus for Chrome
Bytebot
Bytebot
Google Meet Transcripts
Bytebot
Bytebot