Your Own AI Agent, Run For You
A dedicated Grace AI instance for one person, run by us so you never touch a server. Pick a plan by the monthly token allowance you need, or connect your own LLM keys and pay your provider directly. Either way the agent, the updates, and the operations are handled.
Join wishlist
Most AI agents hand your passwords and keys straight to the LLM. Grace AI keeps them in Vault, where the agent can use a secret without ever seeing it.
What Comes With Your Instance
Dedicated Instance
A runtime we run for your account alone. Not a slot in a shared pool where your tasks compete with strangers for the same resources.
Your Keys or Ours
Use the token allowance in your plan, or connect your own LLM keys and pay the provider directly. Switch between the two in settings whenever you like. Keys live in Vault, the agent uses them without the secret reaching the LLM.
Grace AI app
Work from the Grace AI app on your computer, with a mobile app coming soon. The app talks to your instance, not a generic web console.
Token Accounting
Usage is visible per task and per period. Your plan includes a monthly token allowance; beyond it you pay for what you actually use, or run on your own keys instead.
Operations Handled
Backups, patches, monitoring, restoration tests. The runtime care you should not have to schedule your week around.
Privacy By Default
Your instance runs in the EU unless your situation clearly calls for another region. GDPR is a design constraint here, not a line in the footer.
Vault for Your Secrets
Your passwords, keys, and other secrets live in Vault. The agent acts through them on your behalf, the secret never reaches the LLM. You decide what goes in.
No Training on Your Work
Your tasks and files are not Grace AI training data. When you bring your own LLM keys, only that provider's terms apply to those calls.
A Named Operator
Tom J. is the accountable person behind Grace AI. Email or call when something matters. No ticket tiers, no chatbot gate.
Pricing
One person, one instance. The three plans differ by the monthly token allowance included. Prefer to pay your provider directly? Connect your own LLM keys on any plan and the allowance simply sits unused.
Pricing to be announced at launch.
Token allowances and overage rates will be published at launch. Working with a team instead? See Business. Prices exclude VAT.
What Is Included
Everything below is part of your plan. The token allowance scales with the plan you choose; the rest is identical across all three.
- Dedicated Grace AI instance for one user
- Grace AI app
- Vault for your passwords, keys, and secrets
- Your own LLM keys or an included token allowance
- Monthly token allowance per plan
- Encrypted storage, decrypted only when needed
- Scheduled backups with restoration tests
- Runtime monitoring and security patches
- EU hosting by default
- Strict requirements mode for technical tasks
- Direct support, no ticket queue
FAQ
How do the included tokens and my own keys work together?
Run entirely on your own keys, entirely on the included allowance, or mix the two across tasks. You choose in settings. Vault holds whichever credentials you configure. Grace AI bills the runtime and any included-token usage; your provider bills you separately for anything run on your own keys.
What counts as a token?
Tokens for the LLM calls, plus the Grace AI overhead for tools, context management, and orchestration. Exact accounting is documented at launch with per-plan examples.
What happens if I use more than my allowance?
You can let usage run into metered overage at a published rate, move up a plan, or switch that work onto your own LLM keys. Nothing stops mid-task without warning.
What is the Vault?
Vault is where Grace AI keeps your passwords, keys, and other secrets. The agent can act on your behalf using them without the secret ever reaching the LLM. You decide what goes in. Your files are separate: the LLM works with the files you give it, Grace AI just does not push everything to it automatically.
Where does my data live?
On infrastructure we run for your instance, in the EU by default. A specific region is available when there is a concrete reason for it.
Can I move up or down between plans?
Yes. Change plans when your token pattern justifies it. There is no penalty conversation, just a check against how you actually work.
How is this different from a generic chat subscription?
A dedicated instance, Vault, encrypted storage, token transparency, and real requirements on technical tasks. It is built for an AI agent you rely on in real work, not a browser tab for novelty prompts.