Tools and kits for shipping with Claude Code
I build with Claude Code every day and publish what it costs, what breaks and what holds up. Price your own prompts for free, read how the systems work, then take the files.
Start where you are
- I want to know what Claude will cost me Token counter and cost calculator. Free, runs in your browser.
- I want Claude Code to stop forgetting my project Kits with the CLAUDE.md, skills and prompts I use. From $29, paid once. Kits from $29, paid once.
- I want to see how it works before I buy anything 84 posts on Claude Code, ADHD systems and AI search, newest first. 84 posts, newest first.
Chudi Nnorukam. AI systems engineer in the San Francisco Bay Area. Data Science at UC Berkeley.
Price one prompt
Refactor the auth middleware and add tests for the token refresh path.
- Claude Opus 5.5
- $4 in, $20 out per million
- 500 token reply
- 1,000 requests a month
- Prompt size
- 1,240 tokens
- Per request
- $0.0150
- Per month
- $14.96
List prices from Anthropic pricing, verified . An illustration; your own prompt gives your own number.
Published elsewhere
Guides for developers on freeCodeCamp and QuantVPS.
- How to measure your AI citation rate, on freeCodeCamp
- A developer's guide to WebMCP, on freeCodeCamp
- Best LLMs for quant trading, on QuantVPS
What is chudi.dev?
chudi.dev is where Chudi Nnorukam publishes free tools and paid kits for developers who ship with Claude Code. Most of it covers three things: ADHD-friendly Claude workflows, what Claude tokens actually cost, and which model to route a task to.
Frequently asked questions
How do I count Claude tokens and estimate the cost before I send a prompt?
Paste the prompt into the free Claude token counter and cost calculator. It estimates tokens in your browser and prices them on Opus 5.5, Fable 5.1, Sonnet and Haiku. Nothing you paste leaves your machine. For an exact count, use Anthropic's token counting endpoint.
Should I use Opus 5.5 or Fable 5.1?
For most work, start with Opus 5.5. It beat Fable 5.1 on all nine of Anthropic's launch benchmarks at 40% of the price. The full Opus 5.5 vs Fable 5.1 comparison has my measured cost per turn. If you are comparing effort tiers like Fable low vs Opus high, the Fable 5 vs Opus 4.8 tier matrix covers every pairing.
Why does Claude Code say "compaction failed"?
It means your working state lived in the chat, so there was nothing durable for Claude to fall back on. Move that state into files Claude reloads, then run /clear. How I manage Claude Code context across sessions.
Can you build a trading bot with Claude Code?
Yes. I built a 36,000-line production trading bot with it, and task boundaries and verification mattered as much as the model. What building a trading bot with Claude Code taught me.
Are there Claude Code skills built for ADHD?
Yes. I run three: mirror, which counters discounting your own progress with your git log; librarian, which holds the map of a codebase so your working memory does not have to; and standing, which compares your week to your own last month. All three are in a public MIT repo. See the Claude ADHD skills and how to install them.
What is the i-have-adhd skill, and how do I install it?
It is ayghri/i-have-adhd, an open-source MIT plugin by another author. It changes how Claude answers you: action first, numbered steps, no preamble. Install it with claude plugin marketplace add ayghri/i-have-adhd, then claude plugin install i-have-adhd@i-have-adhd, and type /i-have-adhd in any session.
Who is Chudi Nnorukam?
An AI systems engineer in the SF Bay Area who builds Claude Code tooling and AI search systems, and writes up what holds up in real work. Full background and projects are on the about page.
Claude Code AGENTS.md Support: When CLAUDE.md Still Wins