Browse AI
Train a robot to scrape any website in 2 mins with no-code
This was the toughest financial decision I've made in my entire life!
I started doing a trial project with a potential cofounder I met through YC cofounder matching program early September. He told me he tried to go to browse.ai by mistake and found that it's on the expired domains auction.
When I started Browse AI, I wanted this domain, but the owner said he won't go below $50k. So I got BrowseAI.com instead.
The owner said: “Amazon has browse.com, they might want to buy .ai too”
Turns out he forgot to renew!
I talked to my mentor (Daniel Ha, cofounder at Disqus) and he suggested I bid up to 5k for it. I waited until the last hour and right before I start bidding, 2 other people started bidding for up to $250. Then we competed with each other for two hours (the auction was extended) and eventually, I was able to get the domain for $5,009 (plus a 3.627% credit card payment fee). $9 above what the other person bid and I wasn't willing to go any higher...
I'm so glad this happened right before our ProductHunt launch! browse.ai seems much cleaner and easier to read than browseai.com
A few days ago, Browse AI got accepted to
Startup Edmonton pre-accelerator program! As part of this, we may receive up to $50k in equity-free government funding! 💸
I can't wait to get started and see how much value I can get out of it. The orientation is in a week.
I'm going to write a longer post tomorrow about what the process was like.
UPDATE: I wrote a post about the process here:
https://www.indiehackers.com/product/browse-ai/got-accepted-to-a-pre-accelerator--MGll2hK1Opc3-RvTGag
-
1
Wow. Your project is pretty ambitious. Good luck.
-
1
Thank you ☺️
-
-
1
Congrats!
-
1
Thanks Michael!
-
-
1
nice!!
-
2
Hey fellow Canadian! 🇨🇦 👋
-
-
1
Congrats!
-
1
Thanks Tommi! 😀
-
-
1
Great!
-
1
Congratulations !
-
1
Thanks! 😀
-
-
1
No way! I am in the preflight program right now, I saw the list of companies going into the accelerator and yours was the only one I looked into because it stuck out to me. Goodluck!
-
1
That's awesome!! Thanks!
What is Preflight like? I don't know much about it. Has it been useful?-
1
It's very much more focused on the fundamentals, how to talk to customers, how to decide on which customers to focus etc. It's been good so far! Gives a lot of clarity to processes that seem like common sense but aren't necessarily.
-
-
-
1
You go, @ardalan!
-
1
Thank you! 😃
-
-
1
Congratulations. Let us know how it goes!
-
1
Congratulations! Looking forward to the longer post. All the best!
-
1
What's a pre accelerator program? So they take equity for it?
-
1
They don't take equity. It's all gov funding.
Its goal is to prepare you for rapid growth (and fundraising) by the end of the program.
-
-
1
Awesome, Congrats.
-
1
Thanks Shashank!
-
-
1
Congratulations. Eager to read about your experience :)
-
2
Thank you! I just posted about the process here:
https://www.indiehackers.com/post/how-i-got-accepted-to-a-pre-accelerator-getting-up-to-50k-in-funding-without-giving-up-equity-b04f7b0a08-
1
Awesome. Thanks 🙂
-
-
-
1
Congrats @ardalan !
-
1
Thank you Boris!
-
-
1
This comment was deleted 3 years ago
-
1
Yeah for sure! There's a lot but here's the gist:
- React
- Gatsby
- Puppeteer
- AWS Lambda
- Serverless Framework
- 2Captcha for solving captchas
- Proxy servers to avoid getting blocked
At its core, it's a web automation software that can automate almost anything you do on the web, but we're focusing on scraping and monitoring for now.
-
2
This comment was deleted 3 years ago
-
We made our user dashboard and Chrome extension public this week. Anyone can now sign up and use the product. I'm now sending invitation emails to batches of 5-10 people from the waitlist every day, hoping to get feedback.
Even though it took us ~7.5 months to get here, there are still important features we haven't finished developing yet (Capture Text, for example). Moving forward, we'll work on those features while gathering feedback and readjusting priorities accordingly.
Our private launch was scheduled for today.
We knew we were probably going to be a few days late because we're behind on a few bug fixes and features, but I didn't think we'd be blocked by something completely out of our control...
That was until last weekend when I submitted our Chrome extension to Chrome Web Store to be used by Testers and I found out even if you're only launching privately to a few specific accounts, your extension has to go through a manual review process by Chrome Web Store team. 😮 And guess what...they're currently going through the backlog of extensions submitted 3 weeks ago! 😱
I was so pissed at myself for not submitting the extension sooner. 🤦♂️ Because the extension is pretty much done and the changes we're making are mostly on the dashboard and the backend services. I was thinking we may have to postpone the private launch until a month from now...which is super disappointing when you've been working 70 hours/week for many weeks to get your product launched in time.
An hour ago I refreshed my pinned Chrome Web Store developer dashboard tab and saw this 😍
SUCH A HUGE RELIEF!
I guess we're going live now! 😍 Just need to wrap up a few things and start sending the invites to the 60 signed up early users plus friends and ex-coworkers! Exciting times!
-
1
Hey Ardy, we are also developing a Chrome extension and had the same problem. If you are testing internally with whitelisted users, I really do not know what is the rationale of having the extension manually reviewed. It is devastating to wait, if you have just been building it for so long. Took us 3 weeks+ to get the first version on Chrome Web Store, but after that (if you do not change permissions) the review is quite fast - in 1-2 days. Good luck with your product, looks superb!
-
1
Thanks for sharing your experience!
The whole review process is a pretty big hassle. For example, we want to send invitations for our "Testers-only" extension and at the same time submit it for a review to be published on the webstore (Unlisted) so that even people who don't have a Google account can download it through a link. But...if we submit it and it takes weeks to review, during those weeks, we wouldn't be able to publish bug fixes for our testers... so we submitted the extension separately instead of changing the publish settings on the Testers-only one. Now waiting for it to be reviewed.-
1
Congratulations on the launch Ardy! Browse AI fell on my radar a while ago on twitter by someone from the no code crowd. It's a fascinating tool and I'm waiting for a good opportunity to use it.
As @lowraw mentioned, we went through the exact same story with publishing Tablerone on CWS. Their dashboard is a disaster and their review staff is not much better. It took them a month to confirm some of our assets.
The worst thing is that they disabled reviews for everybody. I think after the Tik Tok Play Store fiasco to avoid similar incidents in the future.
-
2
Thanks Natan! Our extension was just approved! We're going to (soft) launch this week.
I hadn't heard of the Tik Tok Play Store incident. But I've heard there are a lot of extensions posting fake reviews and even fake downloads. Given that most apps businesses use have become web-based, I think it's crazy that Google is not investing 100x more in their extension store. There's such an enormous untapped potential there.
-
-
-
We just launched the first version of our site: BrowseAI.com
Previously, there was a simple landing page there that I built back in January to attract some early users. 57 people signed up ever since which is double what I was hoping for (30) mostly thanks to the IH community.
Now, we are hoping to attract more people while we launch privately to our early users in a couple of weeks. We're super excited to hear people's feedback on the product! We also posted on IH to get some feedback on the site:
Why would you 🔥NOT🔥 sign up for Browse AI?
We had 25 early users on the waiting list yesterday. Our goal was to reach 30 before our private beta mid-May.
Indie Hackers tweeted about Browse AI yesterday and that brought 7 more early users in a few hours!
I posted our experience here:
https://www.indiehackers.com/post/our-saas-now-has-32-early-users-thanks-to-a-tweet-by-ih-bff331d5b7
We had 25 early users on the waiting list yesterday. Our goal was to reach 30 before our private beta mid-May.
Indie Hackers tweeted about Browse AI yesterday and that brought 7 more early users in a few hours!
I posted our experience here:
https://www.indiehackers.com/post/our-saas-now-has-32-early-users-thanks-to-a-tweet-by-ih-bff331d5b7
A developer friend of mine told me a few weeks ago he quit his job and is now looking for opportunities, in case I know someone who's hiring. I wasn't planning on hiring a developer before my product has some traction, mainly because I'm bootstrapping the business with my personal savings, but I have a lot of trust in this person, so I decided to take a chance and hire him. However, we decided that he should spend a few weeks on learning about some of the technologies I'm using (Typescript, GraphQL, AWS Lambda, Chrome extension APIs, ...) before he starts.
He started on Feb 25th, and within a few days he was making valuable contributions. Now I feel MUCH more comfortable stepping away from product development sometimes to focus on strategy, sales & marketing, and talking to customers. And I'm much more confident we'll be able to send invitations to our early users before the end of this month.
A few things that helped:
- Using Typescript! He said the typescript types are like documentation to him (even though this is his first TS project). This, and the few weeks he spent on learning, made the onboarding extremely easy and quick even though the project is quite complex from a technical perspective.
- Having enough savings! I worked as a full-time employee for 6 years and saved aggressively. I could have quit earlier, but I wanted to have enough savings to be able to hire people and spend some on marketing if I need to. It wasn't only about covering my living expenses.
- Knowing the person. This took a lot of hiring challenges away. I fully trust him, and I know he is incredibly smart and passionate about what he does.
- Working remotely. I plan to keep the startup remote and not get an office. My friend is in another country and I couldn't hire him if I wasn't open to (and used to) collaborating remotely.
I just published the first landing page for the product on BrowseAI.com . I also posted on IH forums to hopefully get some feedback from others in the community:
https://www.indiehackers.com/post/trying-to-make-web-automation-easy-just-published-the-landing-page-8432dd8655
How is picking a name still one of the most difficult tasks for a startup??
Every good domain is already registered. So it's nearly impossible to find a name that's good AND has a good domain available.
I spent probably 20 hours over 20 days on just the name, which is waaay more than what I was expecting. But eventually I came up with Browse AI which I really like, and registered the .com domain.
A few tips for people stuck on the name like me:
-
Coming up with a name requires a lot of creativity, and if you're like me your ideas could change with your mood! So I recommend allocating small chunks of time during several days to this instead of spending one or two full days on it. You'd be surprised how your ideas improve over a few days and how they vary from day to day.
-
I heard on a podcast (I think it was IH) that YC recommends instantdomainsearch.com for looking up domains. It's the fastest one I've found, but BE CAREFUL! sometimes it shows false positives or negatives. I found Namecheap.com to be quite reliable, but not as fast. With other services like GoDaddy, I've heard a lot of people saying they search a domain there and it gets bought by the provider shortly after. So I wouldn't use them.
-
Instead of pushing yourself to come up with a domain or name first, try to come up with some keywords that you want most associated with your product. Then enter them on namelix.com and go through a hundred results.
-
Enter your keywords on Google like "define [the keyword]" and see what synonyms show up.
-
Document all your keywords and names that have a good domain available. I used Notion for this. Don't rely on your memory. Our memories as human species living in the age of Google really suck.
-
Once you pick a name, don't forget to Google it and make sure there's no other company (at least in the same space or region) using the same name.
5 Comments
Congrats you seem to be tracking well, so I tried it out. mm.. results not so hot. Would be great to see an update here of how your tracking.
Hi there. Could you send me a link to your robot so I can take a look and find out what the issue was?
We have an article on your AI question:
https://help.browse.ai/article/138-leveraging-ai
chatgpt plus reply:
AI (Artificial Intelligence) and machine learning are related but not the same thing.
AI is a broad field of computer science that aims to create intelligent machines that can perform tasks that typically require human intelligence, such as understanding natural language, recognizing images, and making decisions.
Machine learning is a subset of AI that involves teaching machines to learn patterns in data, without being explicitly programmed. Machine learning algorithms are trained on a large amount of data, which enables them to learn from examples and make predictions or decisions based on new data.
In other words, machine learning is a specific technique used within AI to build intelligent systems. It is one of several approaches to building AI, including rule-based systems, evolutionary algorithms, and neural networks.
So while machine learning is a key part of AI, they are not interchangeable terms. AI is a broader field that encompasses many different techniques and approaches, while machine learning is one specific approach to building intelligent systems.
Thanks for the feedback. I think this is more of an opinion though rather than a fact. If we treat this (“ML is a subset of AI and if you only use ML, you can’t say you use AI”) as a fact then you can’t call any application a “software” because what you’re building and the tech you’re using are a tiny tiny subset of what a software could be and the tech you could be using, so you can only call it something a lot more specific.
To be clear, we never claimed that we use “generative” AI anywhere (although that is going to change soon…).
Still, I appreciate the feedback.
re ticket: I have already at the end of the robot session. No support reply recd. Maybe your process is broken.
For the lay, generative AI is not same as ML.
So for me, Browse AI really is Browse ML. misuse of "AI" as it is subset.
Its cool to get domain to "grow into". Uncool to mislead. Suggest being transparent around this if you dont wish to aleniate the people who are willing to try out your extension.
chatgpt plus answer:
Generative AI and machine learning are related but not the same thing.
Machine learning is a type of artificial intelligence that involves training a computer system to learn patterns in data, without being explicitly programmed. The goal of machine learning is to enable the computer system to automatically improve its performance at a specific task, such as recognizing images or understanding speech, by learning from data.
Generative AI, on the other hand, is a subset of machine learning that involves using algorithms to generate new data that resembles the data used to train the model. In other words, generative AI is focused on creating new content, such as images, music, or text, rather than simply recognizing or classifying existing content.
So while generative AI uses machine learning techniques to generate new content, it is not the same thing as machine learning. Machine learning can be used for a variety of tasks beyond generation, such as classification, regression, and reinforcement learning.