What AI coding agents really cost
Field notes from running AI coding agents in production: real bills, real token mixes, what broke and what worked. Written for solo developers and small teams.
New write-ups and model price changes, about once a week. Free. No spam; unsubscribe in one click.
- 64 write-ups from production
- 68 models repriced daily
- Every number sourced
Three tracks
Built with agents
One person, a coding agent, a shipped product: the real build logs, with the hours, the bills and the step that broke.
- Your AI Didn't Misread Your Code by Accident. You Handed It the Wrong Context.
- How to Turn the Workflows You Won't Document Into Agent Skills
- The AI Automation Ceiling: Why 60% Efficiency Doesn't Equal 20% Conversion
Sold
How AI products actually get customers and revenue: teardowns with the numbers and where they came from.
- How a Multi-Agent AI System Made $7,600 in 7 Days for Under $100 in API Costs
- Stop Building Custom AI Agents: How to Earn $63K/Month with Micro-Automation
- 63 Billion Tokens in One Month, Spent Writing Google Ads for Amazon Sellers
Tested
Open models and cheap coding stacks, measured: price, limits, speed and what they do inside a real coding agent.
- How Alibaba’s Open Code Review Slashed AI Code Review Costs by 90% — And What It Means for Independent Developers
- Bypass Codex Rate Limits: The Local Proxy Path to 70% Cost Savings
- Stop Hitting the 5-Hour Limit: Routing Your IDE’s AI Requests to Local Models
Latest write-ups
All 64 write-ups →How a Multi-Agent AI System Made $7,600 in 7 Days for Under $100 in API Costs
An AI crypto trading bot reportedly made $7,600 in a week.” That line gets clicks.
How Alibaba’s Open Code Review Slashed AI Code Review Costs by 90% — And What It Means for Independent Developers
In July 2026, Alibaba open-sourced Open Code Review.
Bypass Codex Rate Limits: The Local Proxy Path to 70% Cost Savings
If you're hitting the 5-hour Codex limit like clockwork every day, stop counting minutes.
Your AI Didn't Misread Your Code by Accident. You Handed It the Wrong Context.
"If your AI keeps writing bad code, you need a better model." In each of the rebuilds below, a new model wouldn't have fixed anything, because the model was reading the wrong thing.
Stop Hitting the 5-Hour Limit: Routing Your IDE’s AI Requests to Local Models
The "5-hour limit" on AI coding agents like Claude Code is the primary bottleneck for independent developers, forcing mid-flow halts just as you hit a rhythm.
Four Circuit Breakers Every Unattended AI Pipeline Needs (Learned the Expensive Way)
An automated game guide pipeline was left running for a month.
Stop Building Custom AI Agents: How to Earn $63K/Month with Micro-Automation
Jordan, an independent developer, now earns $63,000 a month from Resellbot, a tool that handles the daily sharing Poshmark sellers used to do by hand.
I built a WeChat Official Account writing Agent that cuts work time by 80% — here’s the workflow filter and tool combo I used
My WeChat Official Account writing Agent now finishes its run in about 11 minutes of machine time.
The Hidden Costs of Over-Prompting in AI Coding: Lessons from Claude Code's Optimization
The myth that more detailed prompts always lead to better AI coding outcomes is being debunked by developers who have seen firsthand how excessive prompting can actually reduce efficiency.
Browse by topic
Tools and data
Know what your agents will cost before the bill does
New write-ups and model price changes, about once a week. Free. No spam; unsubscribe in one click.