Frequently Asked
Questions

Lexi charges 40% of the savings we generate for you. If your original request would cost $1.00 and STONE restructures the tokens so the provider charges $�.30, you saved $�.70. Lexi fee = $�.28 (40% of $�.70). You pay $�.58 total instead of $1.00.

You pay $� in Lexi fees. If STONE cannot reduce the token count for a particular request, the request passes through at the original provider cost with no additional charge. Zero risk.

Log in to your dashboard and click Add Credits. Payments are processed securely via Stripe. You can add any amount and credits never expire.

Rate limits are based on your lifetime spend and increase automatically. Starter: 10 req/min, Growth ($10+): 30 req/min, Scale ($100+): 60 req/min, Enterprise ($500+): 120 req/min.

Yes. BYOK (Bring Your Own Key) is fully supported. Add your provider API key in the dashboard and requests will be billed directly to your provider account. You still get STONE optimization and only pay the Lexi fee on savings.

STONE stands for Semantic Token Optimization and Natural Encoding. It is our proprietary engine that restructures your prompts and context to use fewer tokens while preserving the full semantic meaning.

Typical savings range from 50% to 95% depending on the use case. Multi-turn conversations and customer support see the highest savings, while novel research queries see lower but still significant reductions.

Minimal impact. In blind benchmark tests, STONE-optimized requests scored 8.4 out of 9.0 compared to unoptimized requests at 9.0. Most users cannot distinguish between optimized and unoptimized responses.

LexiSaaS supports all major providers: OpenAI (GPT-4o, GPT-4o Mini), Anthropic (Claude 3.5 Sonnet, Haiku), Google (Gemini 2.0 Flash, 1.5 Pro), xAI (Grok), DeepSeek (Chat, Coder), and Meta (Llama). New models are added continuously.