# Jordan Feltner, Product Engineer (Solo) at Feltner Labs LLC — read of CodeBudget, June 16 2026

> "Eight years at funded startups, got laid off from a Series B in January, six months into trying to go indie. I build with Next.js and Supabase, I've shipped three side projects that died quietly, and I'm actively looking for a SaaS idea worth the next 18 months of my life."

## How I got here

Someone in the Indie Hackers Slack shared a link with the comment "this site is weirdly honest, worth a look." That's a higher bar than "check out my landing page" so I clicked. I was not searching for AI cost monitoring specifically. I do use the OpenAI API in a side project so the category made sense once I saw it. I came in cold on the idea marketplace angle.

## What I clicked first

The tagline is fine: "Know what your AI coding costs, before your invoice does." I've had a surprise $300 OpenAI bill before so the fear is real. But the thing that actually stopped me was this line in the scoring block: "$-19,582 Year-1 take-home (Fermi)." A product page that leads with a negative income projection for me, the buyer? That's either genuine or it's a very clever trust play. I read the whole page trying to figure out which.

## Where I paused

The "1 in 9 Meaningful-success odds" line. I stared at it for about thirty seconds. Most idea marketplaces show you the upside, not a 1-in-9 failure rate. The honest disclosure block says: "we don't have live customers on this idea yet. We shipped the strategy package; you ship the customer conversations." That is a genuinely unusual thing to put on a page where you're trying to get someone to hand over $99. I don't know if I trust it yet but it made me slow down and actually read the rest instead of bouncing.

## What I distrusted

"Save 90 percent overnight. Switch from GPT-4 to Haiku for a task." That specific claim in the features list felt off. It's the only number on the product-description side of the page and it's doing a lot of work. Ninety percent is accurate math on list price but the framing implies CodeBudget gives you that savings. The tool surfaces the recommendation; you still have to do the engineering work to switch. That's a classic demo-vs-reality gap dressed up as a feature bullet.

Also: the "buyer clarity: 10/10" and "credibility: 9/10" scores are self-assigned by Wishdeal on their own idea. I have no idea what those axes mean. There's a "How scoring works" link but it's not in the text I have here. Scoring yourself a 9/10 on credibility while simultaneously saying you have zero live customers is a contradiction the page doesn't resolve.

The stock-feature list (real-time spend, drill down by project, model recommendations, budget alerts, export, local LLM support) reads like every "AI monitoring" product pitch from 2024. I can name four tools already doing subsets of this. The page doesn't tell me what makes this one different except that someone at Wishdeal scored it 9/10 on uniqueness, which, again, is self-scored.

## What would convince me

I want to see the Fermi math. Not the summary number, the actual assumptions. What does Wishdeal think realistic MRR looks like at month 6 and month 12, what churn rate did they model, what CAC did they assume? I can decide if I agree with the inputs. A negative year-one number is only meaningful if I can see the logic.

I also want one real screener conversation. Not a case study, because there are no live customers. But if the Wishdeal team talked to 10 CTOs at AI-first startups and found 3 who said "I would pay $X/month for this right now," show me that. Even paraphrased. The absence of that is the loudest thing on the page.

## What I'd ask in an email reply

1. The $99 adopt tier includes "working code starter" -- what exactly is that? A Next.js shell with some OpenAI proxy scaffolding, or something that actually integrates with the OpenAI usage API out of the box? Because the gap between those two is about 200 hours of my time.

2. You scored this 2/10 on financial upside and 5/10 on market openness. Those are your two lowest scores. That combination usually means "competitive market with thin margins." Why does this idea make the cut over the other ideas on your site that scored better on both?

3. Have you or anyone on the Wishdeal team tried to sell this to a real buyer? Not built it, just sent a cold email with a mockup. What happened?

## Verdict: on-the-fence

The honesty framing is working on me, which means I can't fully trust my own reaction to it. The idea category is real -- I've felt this pain -- but the page doesn't show me anything that makes this specific execution defensible over building the same thing myself from scratch, which would take maybe a month and cost me nothing but time.

---
*Memo by skeptic persona, generated 2026-06-16. Studio breaks own self-grading loop.*