building a coding agent with no usage limits sharing the whole journey
Joined January 2026
- Tweets10K
- Following1.2K
- Followers3.1K
- Likes10.3K
Pinned Tweet
Imagine paying $200/month for an AI coding agent and still having to worry about how much you use it.
That's the problem I want to solve.
I'm building my own alternative to Claude Code and Codex.
AI compute is not free.
But I'm experimenting with smarter ways to use it
Can unlimited coding agent be economically viable?
I'm going to find out.
How to go from a vibe coder to AI-assisted programmer in 7 steps:
1. Plan first, code later
Prompt: "Don't write code yet. Give me a step by step plan and ask me questions".
Approve it before you let it write any code.
2. One small piece at a time
Never ask "build me the whole app." Build one feature, test it, then move on.
3. Make it explain itself
Prompt: "Explain what you just wrote like I'm a beginner. What could break?"
If you can't follow it, don't ship it.
4. Save every working version
Use Git. Commit after each thing that works. When AI breaks something, you can easily roll back like it never happened.
5. Ask how to test it
Prompt: "How do I verify this works? Give me all the edge cases that could break it. Write tests for it."
Don't just trust it because "it runs".
6. Get a second opinion
Open a new chat, paste the code.
Prompt: "Review this for bugs and security holes."
AI is a better reviewer when it didn't write the code.
7. Understand errors, don't just paste them
Ask "why did this happen?" before asking it to "fix it". Every bug teaches you a free lesson about the project.
Did I miss a step?
2
2
209
"this AI model scored higher on coding benchmarks so it's better at coding"
but when I'm using a coding agent, I care about:
- how many attempts it needs
- how often I have to intervene
- how much it costs to finish
- whether it breaks existing functionality
- whether the solution actually works
a model can score higher on a benchmark and still be worse for my workflow
14
11
454
Let's Connect 馃
I'm Isaac, a developer building my own AI coding agent because I'm tired of usage limits.
If you're already a mutual, say Hi! 馃憢
If you're into:
Claude Code
Codex
Cursor
AI-powered development
Let's connect!
31
26
939
I think the most the most interesting part about building artesian will be finding out how wrong I am.
Right now, I have an architecture in mind.
I have a list of models I want to use.
I have a strategy for reducing inference costs.
What I don't have is proof that the economics work.
So that's my first real milestone:
Prove that unlimited AI coding can be financially sustainable.
Everything else comes after.
6
3
130
Introducing artesian.
There's something genuinely ridiculous about being in the middle of building something you're excited about...
And suddenly being told:
"You've reached your usage limit."
Not because you ran out of ideas.
Not because you're tired.
But because the tool you're paying for decided you're done coding for now.
That frustration is why I started artesian.
I'm trying to build an AI coding agent where your subscriptions doesn't come with an arbitrary usage countdown.
I know the economics are difficult.
I know there's a reason these limits exist.
But I also know there's an enormous difference between accepting a problem and trying to solve it.
I'm choosing the second.
artesian's waitlist is now open.
If you've ever been interrupted right when you were getting somewhere, this one's for you.
3
1
135
I'm 21, I don't have investors, a team, or millions in funding.
But I'm building an AI coding agent to compete in a space dominated by billion-dollar companies.
Why?
Because I'm tired of paying for AI coding tools only to be told I've reached my usage limit.
I know I'm not the only developer frustrated by this.
So instead of complaining about it, I decided to try building something better.
It's called artesian.
An AI coding agent built around one ambitious idea:
What if developers never had to worry about usage limits again?
I have no idea how far I'll get.
But I'm going to build it in public, share the wins, the failures, and everything I learn along the way.
The waitlist is officially open.
If you believe developers deserve better, I'd love to have you along for the journey.
4
5
204
"be careful what MCP servers you connect to your coding agent"
meanwhile people are connecting their agents to:
- GitHub
- databases
- browsers
- Slack
- production infrastructure
and letting them chain actions across all of them.
we went from worrying about AI writing bad code to giving it access to the entire development workflow.
at what point does convenience become a security problem?
Unpopular Opinion:
Claude Code and Codex have a much weaker moat than most developers realize
You can build a coding agent around models you don't own
You can implement your own tools, context management, memory, and execution loops
You can even let users switch model providers
The difficult part is making all those pieces work reliably
But if the model is replaceable and the harness is replaceable, what exactly stops a small open-source project from building something better?
30
19
525
2900 people now馃コ
my favorite part is recognizing someone's profile picture and remembering what they're building before I remember their name
If you're into:
- AI agents
- Software engineering
- SaaS
- Startups
- Building in public
Let's connect馃
What are you building?
58
40
981
"always review AI-generated code before merging"
okay but we're not using AI to:
- write the code
- write the tests
- review the PR
- fix the bugs
- explain what the code does
and then another AI tells us the first AI did a great job.
at what point does this stop being a code review and become two LLMs agreeing with each other?
are we actually verifying anything?
9
7
451
Let's Connect馃
I'm 21, developer, trying to build my own coding agent because I'm tired of usage limits
If you're already a mutual say Hi!馃憢
If you're into:
- Tech
- AI
- SaaS
- Founder
Let's meet each other!
35
33
718
"always make your coding agent write tests"
but if the same AI:
- interprets the requirements
- implements the feature
- write the tests
- decides what the expected output should be
what happens when it misunderstands the requirements?
the implementation can be wrong and the tests can still pass.
we've basically asked the same student to write the exam and the answer key.
am I missing something?
8
9
170
Looking for more people to connect with馃
I'm 21, a software developer working on something ambitious: a coding agent without usage limits
Would love to meet more people who are:
- Building cool stuff
- Obsessed with AI
- Launching SaaS Products
- Turning ideas into Startups
Already following each other? Say hey馃憢
Not mutuals? Let's change that!
15
19
562