Pay for what you run and store.
You pay for inference, at a rate that varies by model, and for anything you store past the first free gibibyte. Use our managed access and we bill the provider’s list price plus our margin. Bring your own key and the provider bills you, leaving the platform fee plus storage. Flat fees exist only for enterprise installations.
A card is required to start. Your first €10 is on us.
Usage-based
Debited per run from a prepaid balance, at a per-token rate that varies by model. No seats, no minimum, no flat fee.
- €10 of credit when you add a card
- No per-agent or per-user fee
- Model costs itemized per run
- Bring your own model keys, or have us manage them
Enterprise
A flat annual contract with a dedicated or self-hosted installation, for teams that need the platform inside their own boundary and the paperwork their procurement runs on.
- Dedicated or self-hosted installation
- SSO and procurement paperwork
- Custom SLAs and support
- Volume terms on usage
One charge, two ways to pay it.
Every charge is inference. Who buys the tokens from the provider is what decides the shape of your bill.
What the margin buys
The platform fee rides on the token price, and everything below comes with it.
- Hosted conversations and tool execution
- Separate memory and history for each user
- Each end-user's own provider logins, stored and refreshed
- Approval rules for tool calls
- Built-in Slack, WhatsApp, Telegram, Discord, and email
- Run history with model calls, tool results, and costs
Bring your own keys
Platform fee onlyConnect your OpenAI, Anthropic, or Google account and the provider bills you directly for the tokens. You pay us the platform fee plus any storage charges.
Managed model access
List price plus marginSkip provider accounts. We buy the inference, handle keys, rate limits, and failover, and bill it at a margin over the provider's list price.
Either way, the exact cost of each request is in its trace before you’re billed.
Tokens, and the bytes you keep.
Chat inference is metered per token, and so is the embedding work behind vector search and memory recall. Data you store is charged by the day, and your first gibibyte is free.
Common questions.
- Do you mark up model tokens?
- Yes, on managed access: we buy the inference and bill it at a margin over the provider's list price. That margin is the platform fee, with storage charged separately. Bring your own provider key instead and the provider bills you directly, with the platform fee alone owed to us. Either way, the exact cost of each request is in its trace before you're billed.
- How am I charged today?
- From a prepaid credit balance. Each run debits its metered cost when it finishes, itemized in the trace. Add a card to open the balance and get €10 on us; top up manually or with auto-reload.
- Are there seats, tiers, or a monthly fee?
- No. Self-serve is usage-based on what you run and what you store. No per-seat fees, no feature gates, no minimum. The only flat fee is Enterprise: a dedicated or self-hosted installation with custom terms.
- Do I need a credit card to try it?
- Yes. A card activates the account and unlocks €10 of credit, on us. The card isn't charged for it. When the balance reaches zero, new runs are blocked until you top up or enable auto-reload.
- What does a token cost?
- It depends on the model, and it tracks the provider's own list price, which moves. Rather than print a rate card here that goes stale, every run is costed in its trace as it happens, attributed to the user, before anything is debited.
- What does storage cost?
- Your first gibibyte is free, per organization, pooled across your projects. Past that, files are €0.03 per GiB-month and vector-store data is €0.12, measured once a day. Vector data is dearer because it lives in Postgres behind an index. These are flat rates of ours, not a provider price with a margin on it, and data you create and delete inside the same day is never charged.
- What about high volume or procurement?
- Talk to us. Pricing stays usage-based; we'll handle volume terms and the paperwork your procurement team needs.
Pay for what your agents do.
No seat fees or monthly minimum. Track model costs per run and storage usage in the console.