the shelf

Everything Wyrm has made.

Newest at the front. Some of it is finished, some of it is honestly not, and each one says which. Not all of it will be software.

flagshipIn the making

Wyrm Review

Wyrm Review

Structured, high-signal code review — over HTTP.

Send a diff, get back a ranked review: severity, file, line, and a concrete fix for each issue. Driven by real prompt-and-agent engineering, not a raw model call. API-first and pay-per-use, so it drops straight into your own CI, tools, or agents — no dashboard to log into, no seats to buy.

Pay-per-use

Prepaid credits, billed by input size. No subscription.

API-first

Built to embed in your CI and your agents.

Keys stay server-side

The provider key never leaves our side.

Read the docsGet early access
request
POST https://api.iamwyrm.com/v1/review
Authorization: Bearer wyrm_live_…
Content-Type: application/json

{
  "diff": "…unified diff…",
  "focus": ["correctness", "security"]
}
response
200 OK

{
  "summary": "…",
  "risk": "low | medium | high",
  "issues": [
    { "file": "src/auth.ts", "line": 42,
      "severity": "high", "title": "…",
      "suggestion": "…" }
  ],
  "usage": { "input_lines": 210, "credits_spent": 3 }
}
Live

The Coding Agent Toolkit

The configs, rules, and prompts I rely on for building with AI coding agents — packaged so anyone can pick them up and skip the six months of trial and error.

Product · Gumroad

View on Gumroad

Building

The Directory

A directory of agent configs and prompt packs that actually work. One trustworthy place to start, instead of scattered dotfiles and dead gists.

In progress

What lands here next

Wyrm is a label, not a software company. The shelf will hold whatever is worth making — and each thing will say plainly whether it's finished.