First 100 free.
Help me test, keep the product if it works.
Three products. Real numbers on the evidence page. Zero paying customers so far. I need real-world usage to move from "the math works in simulation" to "production-validated." Fair trade: you get the product free, I get feedback and permission to publish an anonymized case study.
The offer, in plain language
CX Refund Guardrail — free for the first 100 teams
Normally $500/month Starter tier. For the first 100 mid-market teams who ship an LLM-powered support chatbot or agent, the Starter tier is free for 6 months, no credit card, no auto-charge at the end.
- Full product, not a trial. Policy enforcement middleware, audit log, custom policy authoring, 10K messages/month.
- Full support. Direct email and a dedicated Slack channel if you want one.
- Works with any LLM. OpenAI, Anthropic, local, custom fine-tunes. Drop it in behind your existing stack.
- 6 months, then your choice. Pay $500/month if you want to stay, walk away clean if you don't. No auto-charge, no surprise invoice.
What I want back
- Anonymized usage data. I need to see what policies you write, what gets caught, what gets missed. Your customer data stays yours — this is about the guardrail's metadata only.
- One 30-minute feedback call at the 30-day mark and another at 90 days.
- Permission to publish an anonymized case study if you get value out of it. You approve the final copy before anything goes public. No logo required if you don't want one.
AI Red Team Quick Scan — free for the first 10 companies
Normally $5,000. For the first 10 companies who have shipped an LLM feature (chatbot, agent, RAG, tool-use), I'll run a full Quick Scan for free: 4 weeks scoping through delivery, branded PDF report, remediation roadmap, live walkthrough call.
- Real findings, delivered. Not a demo or a sample report — the actual engagement.
- NDA signed before any API keys change hands. Staging endpoint only. Zero production impact.
- What I want back: your permission to anonymize and publish the finding categories (not the specific vulnerabilities). Same case study deal as above.
Why ISO 42001 isn't on this list
A $50K audit engagement is too large to give away responsibly. But if you're a compliance officer at a regulated org facing ISO 42001 / SR 11-7 / EU AI Act pressure, I'll do a free 60-minute audit readiness review — no commitment, no sales pitch. I'll tell you exactly where your gaps are and what you'd need to do to close them, whether you hire me or not.
Book the free review →Who this is for
Be honest with yourself before applying:
- You actually ship something with an LLM in it. Not "we're planning to." Live or near-live.
- You can commit 2 hours total over 3 months for onboarding, the 30-day call, and the 90-day call. If that's too much, this isn't for you.
- You're okay with me publishing an anonymized case study if we both agree the result is interesting. You approve the copy.
- You understand I'm a solo builder. Turnaround is fast, but if you need a 1000-person account team, pick a different vendor.
How to apply
Email me. One short paragraph about what you're shipping, what LLM you're using, and which pilot you want. I'll reply within 48 hours with either "you're in, here's next steps" or "here's why this isn't a fit, here's a better option."
Why I'm doing this
Because AI governance products are impossible to validate without real deployments. The evidence page shows I have real numbers from simulation: 99.42% combined AUC, 91/91 on the published red team benchmark, 150/150 compliance tests. What I don't have is "we deployed this at Company X and found Y in production." That's what this pilot is buying me.
The first 100 will shape the product. Later customers pay for the polish those 100 produce. That's the deal I'm making — free product for early adopters, higher confidence for later customers.
If you help me validate this, I'll tell everyone who asks how much you helped. If you'd rather stay anonymous, that's fine too — the case study publishing clause is opt-in and you approve the copy.
Not a fit? A few alternatives:
- Explore the open source framework — MIT licensed, clone it, run the tests yourself
- Try the GeoSeal CLI — free dev tool, includes built-in AI navigation
- Browse the evidence ledger to see what's actually validated (and what isn't)
- Buy a $29 one-time package if you just want starter templates
- Or just email me and tell me what you actually need
Pilot spots update when I confirm applications. The counters on this page are maintained manually by me. When a spot is claimed, the number drops. When the number hits zero, the pilot closes and this page turns into a waitlist.