FAQ

Quick answers to the questions we hear most before and after buying RelyAI. For step-by-step setup, follow the links at the bottom of each answer.

Which AI providers are supported?

Three. You can connect OpenAI, Anthropic, and Ollama (self-hosted or Ollama Cloud). You may add several engines and pick one Default that all AI calls use. Popular picks include gpt-4o-mini, claude-haiku-4-5, and llama3.1:8b. See Connect an engine.

Do I need my own API key?

Yes — you bring the AI provider. You set one system key on Settings → Engines & Keys that RelyAI uses for everyone by default. You can also switch on "Allow customers to use their own API key" so customers enter their own key under Account → AI and run on their own quota instead of yours.

Does it work without a license?

No. Every AI feature is gated behind an active RelyAI license. Until your key verifies as Active on this domain, the AI surfaces don't run and admins are prompted to enter a key. See Licensing.

Is the chatbox read-only, or can it act?

Both, and you choose. In its base form the chatbox is a read-only knowledge-base support bot. Turn on Agent chat and it becomes an agent that can take actions through tools — always confirming anything destructive first, and offering Undo (Cmd+Z or an inline button) while the undo window is open. See The chatbox and Agent tools.

Can the AI change things without asking?

No. Anything destructive requires your confirmation before it runs, and most actions are reversible with Undo. Sensitive fields — passwords, card numbers, CVV, OTP, API keys — can never be filled by the agent. Each capability is also a separate feature flag, and several ship off by default so you opt in deliberately.

What data is logged, and can I audit it?

Everything the AI does is recorded. The observability suite gives you a Usage Dashboard, Conversations you can replay, an Audit Viewer of every tool call (with CSV export), Feedback ratings, and a Self-Improve view of failures and weak spots. See Observability.

How are costs controlled?

Three levers. Quotas & Limits sets per-task rate caps (defaults 20/minute, 500/hour) and an optional per-customer daily cap. AI Credits meters usage on plans that use it. And customer-own-key moves a customer's spend onto their own provider account. Local Ollama models have zero per-token cost. See Quotas, limits and credits.

Does it modify AcelleMail core?

No. RelyAI installs as a standard AcelleMail plugin and mounts its surfaces through AcelleMail's own theme — no core files are modified. See Installation.

How are updates delivered?

In-app. From RelyAI → Settings → Integrity & Updates you check for a new version and click Update now for an in-place download-verify-backup-apply update that auto-restores on failure. A manual Upload Plugin path is available too. See Updating.

What happens if my license lapses?

The AI features switch off until you renew and re-activate. The plugin itself keeps running, and your data stays put — re-activate on the License tab to restore AI. For Lifetime licenses, your installed version keeps working forever; only new releases need an active Support & Updates plan. See Licensing.

Related pages

• Licensing — how the license gate works
• Connect an engine — OpenAI, Anthropic, Ollama
• Agent tools — what the agent can do, safely
• Observability — logging and audit
• Quotas, limits and credits — cost control
• Troubleshooting — when something isn't working