Whitepaper

Private inference, below list price

v1.0October 202610 pagesdraft for discussion

Calling AI models from several makers means one account, one credential and one invoice per maker, a request that carries the caller's identity to every host, and the published list price for anyone below negotiating volume. Moth replaces these with one wallet, one prepaid balance and one key. Every call passes one address, the Mothpoint, where its worst-case cost is held before dispatch and every field that identifies the caller is removed. The paper states the design principles, and the system detail lives in the docs.

1. 01Introduction
2. 02Design
3. 03Privacy tiers
4. 04Pricing and billing
5. 05$MOTH
6. 06Risks and limits
7. 07Future work and conclusion

This paper is not an offer to sell, or a solicitation of an offer to buy, $MOTH or any other token or security, in any jurisdiction. $MOTH has not launched, and nothing in it promises a price, a return or a profit.