Real engineers reviewing your AI-built projects.
A 10-category scored review, anchored to a specific commit, with a branded HTML deliverable and a public badge you can embed in your README. Generated by tooling we've refined across our own products. Verified, scored, and signed by a human engineer.
Or skip ahead: buy a Standard Code Review for $2,000.
Four ways in
Start with a Quick Score. Scale when it pays for itself.
Free, scored, public-badge-issuing automated scan up top. Paid reviews when you need the depth. Retainer when you need a real engineer on speed dial.
Quick Score
Free
per repo, 2 per month
- Automated 10-category scan
- Single letter grade A through F + score / 100
- Top 3 takeaway titles
- Embeddable HRC score badge
- Commit-SHA anchored
Standard Code Review
$2,000
per project
- Full HRC v2.0 10-category review
- Up to 25K LoC
- Branded HTML + markdown + findings.json
- Three headline findings + P0/P1/P2/P3 with cost-of-inaction
- 30-min walkthrough call
- 5 business days
- Re-Review at 50% within 6 months
Comprehensive Code Review
$4,000
per project
- Everything in Standard, plus:
- 25K–100K LoC range
- Deployment-context anchors (regulated / multi-tenant / at-scale)
- Executive-summary appendix (board-ready)
- Optional runtime smoke checks
- 60-min architecture walkthrough
Partner Retainer
$599
per month
- Direct Slack access — questions, second opinions
- 1 Standard review per quarter included
- Badge stays fresh perpetually while active
- GitHub App stays installed for ad-hoc re-reviews
- Referral revenue on jobs you pass to HRC
- Cancel anytime
All paid reviews are anchored to a specific commit SHA. The score and badge represent the codebase at that anchor, not the repo as it currently stands. See a sample report or read the published rubric (Field Notes Vol 06).
The Sycophancy Tax
What a hyperscaler can't ship.
OpenAI, Anthropic, and Google could each ship a free LLM-powered code-review tool tomorrow. Some of them eventually will. Here's what their version structurally can't be — four cards below, or the long-form essay version at “The Sycophancy Tax” (Field Notes Vol 07).
An AI lab's product is the model. Ours is the opinion.
Their goal is maximizing model usage, which forces them to be generic, brand-safe, and disclaimer-heavy. Their free code-review tool can't tell a Fortune 500 customer their code is production-ready — so it hedges everything. HRC is vertical for AI-era builders shipping production SaaS. The rubric, the calibration, the tone, the willingness to say “your auth flow is busted” — that's the product.
The methodology isn't the rubric — it's the operating procedure.
Anyone can publish a 10-category list in a blog post. What they won't publish is the 50-page adjudication procedure behind it — the tone rules, the decisions-vs-findings split, the re-review protocol, the score-cap anchors that prevent a B+ from drifting into a B+, the per-category A-tier expectations. That only emerges from running reviews and refining. We've spent that time.
A real engineer signs your report.
Hyperscalers can't scale to human-touch. They ship products at planetary scale; they don't run consulting ladders. At HRC's Standard ($2K) and Comprehensive ($4K) tiers, Daniel personally finalizes the deliverable, runs the walkthrough call, and signs his name to it. The Retainer ($599/mo) is Slack access to that same engineer. That relationship isn't a feature a free tool can replicate.
The calibration data compounds across reviews.
The refund-on-false-positive guarantee on the $300 Findings Reveal is a calibration loop: every adjudicated refund teaches the system and tightens future reviews. Over 100 reviews, the labeled false-positive dataset on AI-generated code becomes proprietary calibration data that isn't in Anthropic's or OpenAI's training set. Our moat compounds while theirs commoditizes.
Hyperscaler tools will define the category. Specialized companies own the opinionated execution within it. That's true for IDEs (Cursor vs Copilot), for payments (Stripe vs banks), for project management (Linear vs Jira). HRC is that layer for AI-era code review.
Why a partner channel
AI made you a 10× builder. We tell you the truth about what you shipped.
Cursor and Claude won't tell you your code has a problem — they're sycophantic by design. Your friends won't tell you, because they don't know either. HRC scores your codebase against a stable 10-category rubric, surfaces three headline findings with cost-of-inaction, and hands you a badge that says where you actually stand.
Big-4 firms charge $25–75K for a comparable-depth review in 6–8 weeks. We charge $2,000 in 5 business days. The reason: we built the tooling once, refined it across our own products, and run it as a tool — not as a 200-page slide deck.
Calibrated score
10 categories, fixed weights, stable across reviews. Comparable across versions and across customers.
Re-reviews count
Score arc (C+ → B → B+) is your traction graph. Fresh code read each time, mechanically compared against prior findings.
Direct access
Slack on Retainer means "is this dumb?" gets a real answer in hours, not days.
Refer-up revenue
Land a job too big for you? Pass it to HRC and earn a percentage of the engagement.
How it works
Score → report → roadmap → badge.
- 1
Submit your repo
Install the HRC GitHub App on the specific repos you want reviewed (one-click). Or upload a zip if you'd rather not grant any GitHub access. Public repos can also be reviewed via URL alone.
- 2
Get scored
Within 5 business days you receive: a branded HTML report, a markdown working copy, a machine-readable findings.json (drop straight into Linear / Jira / GitHub Issues), and an HRC-internal companion. The cover shows your score, your three headline findings, and the commit SHA we reviewed.
- 3
Walk through it
30 minutes (Standard) or 60 minutes (Comprehensive) covering the three headlines, the Decisions Required section, and what to close first. The action list is prioritized P0/P1/P2/P3 with effort estimate and cost-of-inaction on each P0/P1.
- 4
Ship the fixes, re-review
Remediate at your pace. When you're ready, buy a Re-Review for 50% of the original tier. We do a fresh code read at the new commit and produce a delta report — your score moves, your badge refreshes, and you have a story to tell.
Start with a free Quick Score.
Submit a public repo URL, a private repo with a short-lived token, or a zip upload. Get a letter grade and your top 3 takeaways in under 5 minutes. Email-gated.