Skip to content
Saymail is currently in  closed beta — available by invitation only.
Parcourir la documentation
← Retour aux Docs

Saymail Cloud credits

Pay-as-you-go LLM access managed by Saymail, with Zero-Data-Retention routing.

Saymail Cloud is a managed LLM connection. You buy credits on saymail.eu and Saymail routes AI requests through high-quality providers on your behalf, billing against the credits. You don’t need an API key, you don’t need to pick a model, and the providers Saymail uses run with Zero Data Retention — they don’t keep your message text after the request finishes.

It’s the simplest path to the AI features in Saymail, and it’s separate from the €29.99 app license.

When to use it

  • You don’t have an API key with OpenAI/Anthropic/Google and don’t want to set one up.
  • You don’t want to manage which model is “best this month” — Saymail picks.
  • You don’t have a machine that can run a local model, or you want a fallback when your local model is offline.
  • You want a single bill for AI use, paid in one place.

How it shows up in Saymail

Open the LLM Connections panel. There’s an auto-created Saymail Cloud entry. It’s read-only — you can’t change the model or base URL because Saymail manages those choices.

The entry shows:

  • The connection name and a small “managed” badge.
  • A Manage at shop.saymail.eu button that opens the credit shop in your browser.
  • Your current credit balance (refreshed when you open the panel).
  • A Priority field so you can place Saymail Cloud in your fallback chain.

Buying credits

  1. Click Manage at shop.saymail.eu in the LLM Connections panel, or visit shop.saymail.eu directly.
  2. Pick a credit bundle.
  3. Pay through the shop. Credits land on your account within a minute.
  4. Back in Saymail, the balance updates the next time the panel opens.

Credits don’t expire and aren’t tied to a subscription — buy them when you want, use them when you want.

Privacy

Saymail Cloud only routes through providers contractually committed to Zero Data Retention for API traffic. The request body (your email text and Saymail’s classification prompt) is processed and discarded; nothing is used for training, retained for moderation, or stored for replay. Saymail’s own servers log per-request token counts and timestamps so you can see what credits paid for, but not the message contents.

If you’d rather process email entirely on your own hardware, use a local LLM instead.

Combining with BYOLLM

Saymail Cloud is one priority slot in the same LLM Connections panel as your own keys. A common setup:

  • Priority 1: your own OpenAI key (cheapest for you).
  • Priority 2: Saymail Cloud (covers you when the key has hit a quota or your laptop is offline).

Saymail uses the first healthy connection it finds, so the fallback is automatic.