Most SEO engagements end with a spreadsheet of recommendations that somebody else has to build. I build them. That's the whole difference: the person auditing your rendering strategy is the person who ships the fix.
LLM Gateway — the open-source OpenRouter alternative I co-founded — now pulls 57K clicks from 4.02M impressions a year in organic search at an average position of 8.9, and is currently doing 12K+ clicks every 28 days with the curve still climbing.
Organic search is a large part of how it got from a standing start in May 2025 to $798K in Stripe-verified revenue and $70K MRR across 1,219 active subscriptions — no funding, no audience to launch to. $183K of that landed in the last 30 days alone, which is what compounding looks like once the curve finally bends. The revenue is independently verified through TrustMRR, which reads our Stripe account directly rather than taking our word for it, and the open-source core is public at 1,540 stars. The full playbook is in Building LLM Gateway.
Starlink Prices is the same playbook run solo on a much smaller niche: 9.66K clicks from 830K impressions, average position 7.9, from a standing start — evidence the approach isn't a one-off that depended on a hot category.
#How I actually do it
Programmatic SEO that isn't thin. LLM Gateway's models directory — every model with live pricing, context window and capabilities — became one of its largest organic entry points because each page answers a real question with real data. Provider pages, model comparisons and migration guides ("LiteLLM alternative", "switching from OpenRouter") are built as first-class pages, generated from a source of truth rather than spun from a template. Google has been good at spotting doorway pages for years; the only durable version of this is pages a human would bookmark.
AI search visibility, tracked like a channel. A growing share of developer discovery starts in ChatGPT, Claude and Perplexity rather than a search box. On LLM Gateway I track 50 monitored prompts — the questions our buyers actually ask an assistant — and treat position in those answers as a metric with an owner, not a hope.
Distribution you construct, not discover. Every compounding channel LLM Gateway has was deliberately built: shipping as a built-in provider in OpenCode, listings on models.dev and OpenAlternative, the open-source repo itself. Each one keeps working while you sleep.
Technical SEO by someone who reads the framework source. Rendering strategy, Core Web Vitals, structured data, sitemaps, canonicals and crawl budget on Next.js and React apps — backed by real Lighthouse and Search Console numbers rather than vibes. I've contributed PRs to Next.js, so when the framework is the reason you aren't indexed, I know where to look.
#What you typically hire me for
- Standing up a programmatic SEO surface — data model, page templates, internal linking and the generation pipeline — on a product that currently has a marketing site and nothing else.
- Diagnosing why a React or Next.js app doesn't rank: hydration, rendering strategy, duplicate canonicals, and the metadata that never made it to the server response.
- Building AI search visibility tracking, then the content and structured data that gets you cited in the answer.
- Core Web Vitals and page-experience work on codebases where the previous fix was "add more caching".
- A second pair of eyes on an in-house SEO plan before you commit two quarters of engineering to it.
One honest caveat worth setting up front: this compounds rather than spikes. LLM Gateway's first five months looked nearly flat before the curve bent. If you need traffic this quarter, buy ads — and I'll tell you that on the first call rather than sell you a retainer.
If you're sitting on a product with real substance and no organic front door, send me a note.