Do you store prompts or completions?

No. We store usage metadata: model, tokens, latency, status, timestamps, and the tags you send.

How is that different from Helicone?

Helicone can keep full logs. CostSlice does not. We are not a gateway.

OpenAI cost per customer without storing prompts

Security asked whether the meter stores what people type. We do not. We keep counts, tags, and cost. Bodies never hit our disk.

A security review will ask whether the meter keeps what people typed. CostSlice does not. Prompt and completion bodies are not written to disk. We keep tagged usage: model, tokens, latency, status, timestamps, x-cs-tenant, x-cs-feature, x-cs-env.

Helicone can keep full logs. That is a gateway choice. LLMeter says it is not a proxy — it wraps the SDK in your process. CostSlice is a pass-through at https://api.costslice.com/v1. Completions still go to OpenAI. We never store the OpenAI API key.

Privacy. vs Helicone.

Starter is free. No credit card. Operated by Operator Media Yard LLC, New Jersey. hello@costslice.com

Get started See the P&L

Questions

Do you store prompts or completions?: No. We store usage metadata: model, tokens, latency, status, timestamps, and the tags you send.
How is that different from Helicone?: Helicone can keep full logs. CostSlice does not. We are not a gateway.
How is that different from LLMeter?: LLMeter is a wrapper, not a proxy. CostSlice is a pass-through at https://api.costslice.com/v1. Completions still go to OpenAI.