1/24. @SoftgenAI
- works well for both coders and non coders
- can fix its own bugs
- can host the web apps
It worked really well for my little project (Vacation calendar manager)
2. Cursor
- a fork of a VC code
- raised loads of funding turning into a unicorn
- founders went on a Lex Fridman pod
- turns coders into 10x devs. Not the best fit for non-coders.
3. Wrapifai
- perfect for mini tools (lead magnets or tools to drive SEO traffic)
- mostly produces functioning app right away with one prompt
- doesn't handle serious apps
- unlimited apps (instead of per token..)
4. Windsurf (works well for advanced apps. my top 3)
- acts as a true agent
- offers deepseek
- can use the web search
- has a memory.
5. GitHub Copilot ( the OG of this game, started back in 2020 )
- can generate code
- handles large codebases
- can merge PRs, fix bugs, search code
6. Lovable
- fastest-growing EU based startup today ($10M ARR)
- has native supabase integration
- has the best AI+NoCode mix on the market
7. Bolt
- started as a side project of stackblitz and went huge raising $105M
- works similar to softgen, lovable & windsurf
I built several lead magnets using bolt
8. v0
- best for making well-designed web pages or UIs
- doesn't work well for fullstack apps
- can use figma as a starting point
- lets you edit elements one by one
9. Replit
- most advanced AI Agent for coding imho
- has two modes: Agent or Assistant
- true full stack app generator. Has its own server, db, hosting...
10. MarsX (my own product)
- a mix of AI, NoCode and High Code
- I built @seobotai, @indexrusher and all my other products using marsx
- it basically takes the whole coding world to the next level since it isn't a website builder, but "SaaS builder"
11. Claude
- it can write and run code
- super easy to get started (free)
- works for basic cases. e.g. building components or learning coding
12. Amazon Q
- very few people use it
- totally lagging behind the top players
13. Pear AI
- entered into YC
- had a huge drama around forking a github repo
- had a mentor call with them, two very talented guys taking an alternative path on competing with cursor
14. Devin
- super expensive
- targets corporate world
- acts as a junior dev in a dev team
very good review:
15. Github Spark
- works well for small or demo apps
- still behind the "waitlist"
16. IDX
- free alternative to cursor
- can build mobile apps
- uses gemini
17. Webdraw
- best for people with zero coding experience
- totally genius UX
- free
- turns sketches into web apps
18. Tempo Labs
- Generates full-stack apps using a text or image prompt
- It starts by making an architecture and diagrams
- I think they have the best AI Coding UX, wow
19. Cline
- VC code plugin
- works for large codebases
- supports any LLM
- runtime awareness
20. Continue dot Dev.
- an open source alternative to Cursor
21. Databutton
- from Norway
- backed by VCs who backed one of my startups
- has very unique approach, different from most of the players
- works great for true nocoders
22. Base44
- for noncoders
- all in one
- creates dashboard-like apps pretty well
23. Qodo
- for coders
- can write tests, refactor and generate code
- supports all llms(including deepseek)
24. Caffeine AI
- new player
- too early to say anything. for now just putting them into the list to come back to them later when they launch
25. Aider
- a terminal app for generating web apps out of prompts
I run the most automated org on earth,
thx to the agents I built myself or third parties.
I just saw YC betting it all on "replace humans with AI" ideas.
24 AI agents to automate the entire org:
1. SEObot - automated entire blog SEO.
- figures out my topics, keywords, knowledge tree, and more
- pitches me article ideas daily over an email
- generates them & syncs with my CMS
- does all internal and external linking... -> @seobotai
2. SoftGen - AI Agent for coding.
I build small projects just by talking to it. No coding at all. I don't even read the code. So it basically works for non-coders.
In the old days, I'd give these projects to junior devs & wait for months.
Real demo:
An idiot admires complexity. A genius admires simplicity.
Complexity isn't an asset but a liability. All my simplicity hacks:
1. Solo teams.
One person - one task.
No meetings, no discussions, no arguments, no drama.
2. Google Sheets.
The simplest & most powerful human invention.
I use it for accounting, planning, documentation, backend, and database. It's easy to use, share, edit, clone, automate (this part is crazy advanced, you can automate entire job using google scripts).
3. Less features.
I don't copy an existing tool & add more features on top of it.
Most products are difficult to use because they have too many features for too many audiences.
I pick one audience and remove everything that is irrelevant. Often, that's 90% of the stuff gone.
I started in 2004, exited, fundraised many millions, top accelerators, and then bootstrapped a HoldCo(24 products) to a multimillion profitable business.
7 steps I repeat in every product: Setup>Idea>Build>Grow>Monetize>Hire>Scale.
Full Tutorial:
1. Setup
I had a paid job for the first 5y, to pay the bills.
99% chance you'll fail on 1st try.
Those YouTube stories of 18-year-old dudes getting rich overnight are total bullsh*t cuz they are too rare. It takes 3-7 years. Be prepared, both mentally and financially.
2.1. Idea. First, see it as a training camp.
It isn't easy to come up with good ideas at the beginning. My first idea was a game for coders, in which they code their bot and send it to fight. The second idea was a virtual reality room for team meetings. Both ideas failed miserably because I had no idea what I was doing.
You should not expect yourself to come up with good ideas. Your first 10 ideas will be crap. That's okay. If you have ever played sports, your first 10 tries are awkward; you play it like a toddler making first steps. Why would you expect it to be different in business?
So, your first 5-10 ideas are for you to practice and learn the startup craft.
Back in 2022, I said I'd build the most efficient startup factory in the world.
A "Maker Label" for bootstrapped solo founders.
In 2023, I proved the model: 6 startups, over 100k users & 7-fig in revenue.
My end game was to replace 100% of non-creative labor with AI Agents π§΅
2. My mission was to move the wealth from a closed club
(corporation & VC-backed) to bootstrappers worldwide.
By leveling the playing field:
> making it really cheap to do startups
> removing the "money & VC network" as gatekeepers
> let founders focus on creative tasks
3. I mapped down my entire startup org and tasks to see what's actually being done daily/weekly/monthly.
Then, I started automating it all with AI.
While most builders were focused on AI Wrappers, I went after the AI Agents with human feedback. Similar to a real employee.
[1] Idea.
I only build products that I would use myself. It took me many years of failed projects to stop building stuff for the users I donβt deeply understand.
So the idea = my pain.
[2] SEO potential.
I spend hours on Google search console, digging into keyword traffic trends.
This is the single best way to see if anyone is actually searching for this solution/problem. Very often, the results might be the opposite of what my intuition says. I never skip this step.