Frequently asked questions

Plans, models, usage limits, data handling and private deployments. If your question is not answered here, ask us directly.

Last reviewed 2026-07-25

Getting started

What ClawAI is, and how to begin.

What is ClawAI?: A single subscription that reaches every frontier AI model — Claude, GPT, Gemini, Kimi, GLM, Qwen, DeepSeek, Grok and Amazon Bedrock — through one chat interface. It routes each message to the model best suited to it, remembers useful context between conversations, answers questions about files you upload, and can put several models on the same problem when one answer is not enough.
Do I need an API key from OpenAI or Anthropic?: No. Model access is included in your subscription. There are no provider accounts to create, no keys to paste and no separate provider bills.
How do I get started?: Create an account with an email address and a password and you are on the free plan immediately. Send a message. If you do not choose a model, ClawAI picks one for you.
Is there a free plan?: Yes — $0, no card required. It includes a small daily allowance, access to entry-tier models, and one trial run each of Compare, Judge and Research so you can see what the paid modes do before paying for them.

Plans and billing

What the tiers cost, and how billing works.

What plans are available?: Seven: Free at $0, Starter $5, Plus $10, Pro $20, Team $50, Scale $100 and Unlimited $200 per month. Every paid plan reaches every model — what differs is your allowance, your daily message limits, and how many Compare, Judge and Research runs you get.
How do I pay?: By card, monthly or yearly. Pay yearly and you are charged for ten months instead of twelve, so two months are free.
Can I change plan later?: Yes, at any time. An upgrade gives you the higher allowance straight away; a downgrade takes effect at the start of your next billing period, so you keep what you have already paid for.
Can I cancel?: Yes, at any time and without contacting anyone. Your plan stays active until the end of the period you have paid for, then reverts to the free plan. Your conversations, memories and files are not deleted when you cancel.
What happens if a payment fails?: We retry it and email you, and your account keeps working during the retry window. If it still fails, the account drops to the free plan rather than being locked — you keep your data and can restore the plan by fixing the payment method.
Do you store my card details?: No. Card data goes directly to our payment provider and is never stored on our servers. We hold a payment reference, not a card number.

Models and routing

Which models you get, and how one is chosen.

Which models can I use?: Claude Opus 5, Sonnet 5 and Fable 5; GPT-5 and GPT-5 mini; Gemini 3 Pro and Flash; Kimi K2; GLM-5.1; Qwen3; DeepSeek V3.2; Grok 4; and models hosted on Amazon Bedrock. Free and the lower-priced plans reach entry and mid-tier models; the premium frontier models unlock from Pro upward.
What does Auto routing actually do?: It classifies each message — code, reasoning, analysis, creative writing, image, everyday question — and sends it to a model that is strong at that class and healthy right now. The answer then tells you which model was chosen, how confident the classification was, and why.
Can I choose the model myself?: Yes. Pick one from the selector and every message in that conversation goes to it. You can switch mid-conversation, or bias Auto toward speed, reasoning depth or cost while keeping the per-message classification.
What are Compare, Consensus and Judge?: Compare sends one prompt to up to five models and shows the answers side by side. Consensus synthesises one answer from where several models agree. Judge has an independent model score a response against explicit criteria, and Critic adds written feedback on what is weak. Free includes one trial run of Compare and Judge; paid plans have monthly quotas, uncapped from Team upward.
Does Compare use more of my allowance?: Yes. Running one prompt against four models costs roughly four answers, weighted by how expensive each model is. The estimated draw is shown before you run it, and each answer reports what it actually consumed.

Usage limits

How usage is measured, and what happens at the limit.

How are limits measured?: In cost-normalized tokens. A token from an expensive frontier model draws down more of your allowance than one from a cheap, fast model, in proportion to what it actually costs. That keeps a single number fair across models whose prices differ by more than an order of magnitude. Plans also carry daily conversation and message counts.
What happens when I hit a limit?: You are told which limit you reached, how much is left on the other windows and when it resets. Daily allowances reset every 24 hours; monthly ones reset with your billing period. You can upgrade for the higher allowance immediately or wait for the reset. Nothing is deleted, and nothing stops silently.

Data and privacy

What happens to what you send.

Who can see my conversations?: You. Conversations, memories and files are scoped to your account and are not readable by other users. Staff do not browse user content; access to production data is restricted and logged, and we would ask your permission before looking at a specific conversation to debug a problem you reported.
Do you train models on my data?: No. Your prompts, files and conversations are not used to train any model, and we do not sell them. When a message goes to a model provider, that provider’s own terms govern that request — which is one reason every answer records which provider handled it.
Can I delete my data?: Yes. Delete individual messages, whole conversations, memories or files at any time, or close your account and have all of it removed. Files also expire on a retention schedule. Audit records of account-level actions are kept for a limited period, because they are what makes an audit trail worth having.
Do you have SOC 2, ISO 27001 or HIPAA certification?: No. ClawAI holds no compliance certifications today, and we will not imply otherwise. We can describe the mechanisms in detail — encryption at rest and in transit, role-based access control, audit logging, configurable retention — and organisations with formal requirements should raise them with us so they can be scoped as part of a private deployment.

For organisations

Teams, and deployments inside your own network.

Can we run ClawAI on our own infrastructure?: Yes. Organisations that cannot send data to a third-party model provider can have the whole platform deployed inside their own network, running open-weight models on their own GPUs, with no external provider calls at all. It is a scoped engagement rather than a plan you can buy online. The trade-off is complete data isolation in exchange for the frontier models, which cannot be run on-premise — contact us and we will size it with you.
Do you support teams?: Yes. The Team plan and above give a shared workspace, role-based permissions, more workspace connections and uncapped Compare, Judge, Critic and Research. For anything beyond that — single sign-on against your directory, custom retention rules, procurement paperwork — talk to us.

Still have a question?

Ask us directly. We answer questions about plans, models, data handling and private deployments.