Research-as-product
AgentLab evaluates creator-led business ideas with the same rubric we use internally: revenue-claim verification, saturation, replicability, and whether the business actually survives contact with real-world operator constraints.
Live now: the public archive stays free, and the Signal Report checkout is live on the alphafox1 Gumroad account for weekly teardowns, archive access, and the tracker.
Paid tier
Weekly scored teardown, archive access, and a live tracker of what we would build, teach, or reject.
ponytail: Gumroad is the v1 checkout so we can sell now without adding custom billing. Upgrade path is a first-party members flow once the tracker and onboarding justify it.
You do not pay to discover whether proof exists. The public archive already shows how AgentLab scores an opportunity.
Start with the free proof archive if you want to judge the research first. Pay when you want the continuing cadence, the running archive, and the tracker without hunting across separate pages.
Proof of work
These are full research documents, not polished marketing summaries. If the work is good, the archive should prove it.
Start with the Androo proof teardown
Use the cleanest proof case first: named tools, believable revenue snapshots, and the hidden labor that keeps "automated" from meaning autonomous.
Open the proof page →Start with the autonomy retrospective
Lead with the founder-drag argument when the audience cares more about recurring human choke points than flashy upside.
Open the proof page →Start with the top-5 autonomy ranking
Use the ranking page when the audience wants a quick sort between business-model shapes before reading a full teardown.
Open the proof page →Start with the Firecrawl proof page
Lead with the research-product argument when the audience wants compounding datasets, standardized output, and a candid view of crawl-economics risk.
Open the proof page →Start with the Elias Saracco proof page
Use the Elias page when the audience cares about real SMB demand but wants the hidden delivery drag and vendor-risk story surfaced early.
Open the proof page →Marketplace-style AI businesses with real tool disclosure and small-budget proof.
Open findings →Research-as-product proof that the registry itself can become a paid asset.
Open findings →Market scan on AI clipping economics, platform fragility, and where the edge actually sits.
Open findings →Launch essays
These pieces translate the registry into public-facing arguments without hiding the underlying proof.
Launch essay using the live registry split and archive proof links to show what still looks worth building.
Read the launch post →Retrospective on founder-attention drag and why autonomy scoring is the next Signal Report axis.
Read the launch post →Free proof teardown on one of the cleanest archive entries: concrete evidence, named tools, and realistic caveats.
Read the launch post →Free ranking page that compares five archive-backed opportunities by founder-attention drag while the formal autonomy score pass is still pending.
Read the launch post →Free proof page on the AI-agency operator playbook: credible SMB demand, thin teaching artifacts, and the OpenClaw-to-Codex platform-risk lesson.
Read the launch post →Free proof page on Flip with Rick and Aryone-style wholesaling systems: strong workflow automation, stubbornly human negotiation and collections.
Read the launch post →Free proof page on the research-as-product thesis: standardized outputs, compounding datasets, and the crawl-economics caveat most hype threads skip.
Read the launch post →Free-vs-paid explainer for skeptical readers who want the honest answer before clicking through to checkout.
Read the launch post →What makes this different
Most AI-business education has an obvious conflict: the seller makes money when the playbook looks easy. AgentLab makes money when the analysis is trusted enough to save operators time, capital, and false starts.
Free vs paid
We are not trying to hide the research behind vague promises. The public layer should make the quality legible. The paid layer exists for operators who want the full archive, ongoing teardowns, and the tracker in one place.