Clarity. Rigor. Precision.
AI tools for sharper thinking.
Foretoken Labs builds AI tools designed to supplement and sharpen human thinking — not to replace it. We are experimenting with how AI can help people reason more clearly, and produce the best products possible.
Available now
Steel-Manning Workbench
Our first product. We started here because steel-manning — stress-testing an argument against its strongest opposition — is one of the most demanding thinking tasks there is, and one where AI can add real, traceable value.
You submit a document, confirm what the system extracted, and receive a four-part analysis: where your argument is weakest, the strongest case against it, and what to do about both.
Argument Map
Your primary claim, supporting claims, assumptions, and evidence — confirmed by you before the analysis runs.
Vulnerability Map
Where the argument is weakest, ranked by severity.
The Steel Man
The strongest possible opposing argument — including the ones your adversary hasn't made yet.
Threat Assessment
Specific, traceable recommendations for hardening the argument against each high-severity threat.
The difference
It uses your inputs, not its guesses
Before any analysis runs, you confirm the claim, assumptions, domain, and likely adversary. The output is anchored to what you actually meant.
It refuses to ship a weak result
An independent model reviews every steel man. If it isn't strong enough, the system regenerates. If it still can't pass evaluation after regeneration, you get a failure notice and pay nothing.
It applies a real method, every time
The pipeline encodes established frameworks from argumentation theory — Toulmin structure, the RSA framework, the MAFALDA fallacy taxonomy — applied consistently. You do not need to know the theory. The product does.
Ready to pressure-test your argument?
Start with 3 free analyses — no subscription required.
Get Started