A provocation is a bold, overdrawn claim about a phenomenon existing theory fails to explain. State it at maximum force — the sparring partner will push back. Don't soften it yet. The challenge is the method.
Extrafemi Labs never sees your research. This tool is a static HTML file that runs entirely in your browser. There is no Extrafemi Labs server, no database, and no logging. Your conversations never touch any infrastructure we control.
Anthropic and OpenRouter process your API requests — this is unavoidable given how the tool works. Neither provider uses API-tier data to train models by default. See their policies: Anthropic Privacy Policy ↗ · OpenRouter Privacy Policy ↗
Connecting here unlocks "Begin sharpening" on step ④. Your provocation, domain, and connection are passed directly to the tool.
Whetstone is a static HTML tool. There is no server, no database, no user accounts, and no logging infrastructure operated by Whetstone or Extrafemi Labs. All processing happens in your browser.
Your API key is stored only in your browser tab's memory. It is never sent to Whetstone, never written to local storage, and disappears when you close the tab. You are responsible for the security of your key.
Everything you type is sent directly from your browser to your chosen AI provider (Anthropic or OpenRouter) via their API. Whetstone does not receive, process, or retain any of this data. Anthropic Privacy Policy: anthropic.com/legal/privacy · OpenRouter Privacy Policy: openrouter.ai/privacy
The Whetstone Disclosure Report is a record of session activity generated at export time. The SHA-256 hash certifies the document has not been modified since export. It does not constitute legal evidence or a guarantee of academic integrity. Interpretation remains with the evaluating party.
Whetstone is provided as-is with no warranty of any kind. AI-generated content may be inaccurate, outdated, or fabricated. Always verify citations independently before submitting or publishing. By using this tool you confirm you are 18 years of age or older.
These terms are governed by the laws of the Federal Republic of Nigeria.
Whetstone gives every researcher access to the intellectual practices that elite institutions have always reserved for their own — the seminar culture, the hard pushback, the literature that knows your framing is twenty years behind. What survives is defensible.
High-quality research has never required access to information. The internet solved that years ago.
What it has always required — and what has never been equally available — is access to the intellectual practices that turn information into a defensible contribution: adversarial debate, red teaming, steel-manning.
Those practices have historically lived within elite institutions. Seminar cultures. Senior colleagues willing to push hard. The informal intellectual life of well-resourced departments. If you were not in proximity to that environment, you were working without the sharpest available tools, regardless of how good your ideas were.
Now, that problem is gone. Not because AI is a genius. Because AI can be configured to behave the way a rigorous intellectual environment behaves — and because a generation of academic publishing has been fed to it. It has breadth. It knows the literature.
Whetstone is that configuration.
The intellectual practices research has always required — priced, for the first time, at what they actually cost to run.
This is not a metaphor. Academic publishing houses licensed their peer-reviewed content — previously justified as requiring paywall protection — to train AI systems. That research is now inside the model. Whetstone deploys that breadth as an adversarial instrument calibrated to your domain and literature base.
Most tools that involve AI and research treat AI as an author, an assistant, or a teacher. Whetstone is none of these. The critical distinction is when in the research process challenge is applied — and who the AI is in relation to the researcher.
| Framework | AI role | Pre-formation challenge | Citation enforcement | Contribution record | Paper production |
|---|---|---|---|---|---|
| Whetstone — IAI v3.0 | Adversarial sparring partner | ✓ Stages 1–4 | ✓ 3-tier enforced | ✓ Full session record | ✓ Track 2, Stages 5–8 |
| Hypothetico-deductive method | None | ✗ Post-formation only | ✗ | ✗ | Standard peer review |
| Socratic method tools (Kialo, etc.) | None or minimal | Partial — user-driven | ✗ | ✗ | ✗ |
| Standard peer review | None | ✗ Post-formation only | Varies by journal | ✗ | ✓ Publication standard |
| AI writing assistants (ChatGPT, etc.) | Author / co-author | ✗ | ✗ | ✗ | Draft generation only |
The critical difference: in hypothetico-deductive research and standard peer review, challenge is post-formation — the thesis is formed before it is tested. IAI applies challenge pre-formation, at the point where it is most generative: the researcher can reconceive rather than merely revise.