The three free starter kits are a good baseline. If your stack is weird — polyglot, regulated, private-SDK-heavy, or specific in a way no generic kit covers — commission a hand-built Claude Code / Cursor / Copilot config. Pick a tier, pay once in BTC to the listed address, paste the tx hash, compose and send the brief email, then receive a private custom AI coding config ZIP within 48 hours.
Choose Starter, Standard, or Deep below. Each tier shows the exact sat amount, expected file count, included revision rounds, and what kind of private custom AI coding config ZIP you get.
Send the listed tier amount once to the BTC address on this page. Any confirmed Bitcoin Layer 1 transaction counts — we verify on mempool.space before starting.
After paying, paste the transaction hash into the form. "Compose email" fills a message with your tier, tx hash, reply email, and stack brief; you review it and press send.
Within 48 hours we reply with a private download link to your custom AI coding config ZIP. Revision rounds included — one for Starter, two for Standard, three for Deep. If we can't deliver, we refund in BTC to a refund address you provide.
USD estimates above are approximate and recomputed live from the Coinbase BTC-USD spot API when you load this page. If the API is unavailable, the static approximation shown is our best fair estimate at time of writing. The BTC amount in sats is the authoritative price — USD figures are informational.
CLAUDE.md skeleton, 20 power prompts, slash-commands, hook scaffolding. Best if your main AI is Claude Code.
Kit 002 · free.cursorrules templates, agent-mode directives, tab-vs-composer guidance. Best if your main AI is Cursor.
Kit 003 · freecopilot-instructions.md, MCP config stubs, language-specific prompt files. Best if your main AI is GitHub Copilot.
If a free kit already covers 80% of what you need, you probably don't need a commission. Commission tier value-add is the custom 20% — the part nobody would ship as a generic template because it's specific to your codebase.
Primary programming language(s), frameworks, build system, the AI coding tool you use (Claude Code / Cursor / Copilot), and one or two workflow pains you hit regularly. Optional: team size, monorepo layout, private SDKs you want the AI to respect. If it's in the brief, the kit will account for it.
If we can't deliver a kit within 72 hours of a confirmed transaction (48h turnaround + 24h buffer), or if the kit is clearly off-scope compared to your brief, you get a full BTC refund to any address you provide. Refunds after delivery-and-acceptance are partial and scoped — if the revision rounds don't resolve a miss, we refund the portion unearned (typically 20–40% depending on how much of the kit landed right on first pass).
Everything the three free kits already cover — JavaScript/TypeScript, Python, Go, Rust, Java, C#, Ruby, PHP, Swift, Kotlin — plus SQL dialects, infra-as-code (Terraform, Pulumi, Kubernetes YAML), and front-end frameworks (React, Vue, Svelte, Astro, Next.js, Nuxt, SvelteKit). For very niche stacks (Elixir-Phoenix, Clojure, OCaml, Zig, etc.) we'll still take the commission but may downgrade the tier if coverage thins out. You can ask before sending BTC; we'll reply within 24h with a "yes we can cover this" or "no — here's why" note.
One round for Starter, two for Standard, three for Deep. A "round" means you reply with specific corrections or additions; we ship an updated ZIP within 24h. Revisions beyond the included count are priced at half the original tier — e.g. an extra Standard round is 75k sats.
48 hours from the moment your transaction hits at least one mempool confirmation AND you've sent the email with your brief. If BTC lands late-Friday we may roll into early-Monday morning — the 48h clock pauses across weekends only if you specifically allow it. Otherwise it runs continuously.
Both. Drafts are AI-authored under human direction; every file is reviewed and edited by a human before delivery. We are upfront about this because AI-generated configs for AI tools are not a neutral choice — we use the tools we're configuring, so we catch their own quirks. If you want a 100%-human-authored kit, ask before sending BTC; the Deep tier is the only tier where that's feasible, and it'll be a 72h turnaround instead of 48h.
Bitcoin mempool latency is your risk. If you underpay the network fee and the tx sits unconfirmed for >24h, the 48h clock hasn't started. We recommend using a fee-estimator (mempool.space's next-block rate is a good target). Once the tx has any confirmation, the clock starts.
BTC Layer 1 only for now. It's the only payment rail with zero onboarding cost on our side. Lightning may be supported in future; for now, on-chain BTC is the single path.
Nowhere except your email client. The page composes a mailto: URL with the tier, brief, reply email, and tx-hash as prefilled subject and body. Your email client opens; you click send. Nothing is uploaded by this page, and the reply address is a plus-aliased inbox we filter on.
Just browsing? Tip the maker in BTC — no account, no signup, just paste.
bc1qs04leape97ner4wqa98n94l9n0gv9aa84eg4ux
Hand-built single-file games, quizzes, and visualizers — no signup, no tracking, no cost.