Terms of use
Effective September 29, 2026
What FactCan is
FactCan analyzes a short claim you submit and returns a structured response. Depending on the engine mode this release is running, that response may include model-generated Truth Tiles with a verdict and falsifier, or it may state that generated tile analysis is not available and return only classifier and (if enabled) retrieval signals. The current mode is shown on the FactCan page.
No professional advice or warranty
The service is experimental and provided as available. Any generated interpretation, classifier score, or retrieval match may be incomplete, wrong, or misleading. A quotation located during retrieval does not establish that every claim in the analysis is true, and it is not independent fact-checking. Honeycutt AI Labs LLC does not warrant uninterrupted operation, accuracy, or fitness for a particular purpose.
Acceptable use
Submit only claim text you are entitled to share for analysis. Do not use FactCan to generate or launder claims that misrepresent people, and do not submit content you know to be illegal. Requests over 2,500 characters are rejected; FactCan is scoped to short, viral-post-length claims, not long-form articles.
Capacity
FactCan queues each claim on a single shared host and analyzes one claim at a time on a Honeycutt Ai Labs-owned machine. The queue is small, claims are kept for at most one hour, and a full queue or a used-up daily budget returns an error asking you to retry. When the generation machine is offline the page says so and returns only the classifier signal. See the privacy notice for how a queued claim is handled.
Changes
Material changes will be posted here with a new effective date.