Frequently asked questions

Kadeon offers a Free plan (with monthly limits on tokens, documents and storage), a Pay-per-use plan (you pay only what you use, billed at end of cycle) and Subscription plans at a fixed monthly, quarterly or annual rate. See the Pricing page for up-to-date details.

Yes. The Free plan is completely free and requires no payment details. You can sign up and start using the service immediately.

The Free plan has monthly quotas on AI tokens, number of processed documents and storage. When you reach a limit, some features are temporarily blocked until the next cycle starts. You can always upgrade to get higher limits.

Tokens are the unit of measurement for AI model consumption. Each word processed by the AI (prompt + response) corresponds to approximately 1–2 tokens. The amount of tokens consumed depends on document length and extraction instructions.

For Subscription plans: at subscription and then automatically at each renewal (monthly, quarterly or annual). For the Pay-per-use plan: at end of billing cycle, based on actual consumption of tokens, documents and storage.

Payments are handled through Stripe, which supports credit/debit cards (Visa, Mastercard, Amex) and other local methods. Card details are never stored on our servers.

Yes. Stripe automatically issues a receipt/invoice for each payment to your account email address. You can view and download all invoices from the Stripe portal accessible from Billing → Manage subscription.

Stripe will automatically retry in the following hours/days. If the payment remains unpaid, the subscription may be suspended and your account will return to the Free plan until resolved.

Go to Billing and choose the desired plan. You'll be redirected to Stripe to complete payment. The upgrade is immediate.

Yes. First cancel the current subscription (through the Stripe portal) then activate the Pay-per-use plan by adding a payment method. The two plans cannot be active simultaneously.

Yes, all data, extracted records and configurations are kept unchanged during any plan change.

If you already have an active Stripe subscription, when upgrading to a higher plan Stripe automatically calculates the credit for unused days in the current cycle and charges you only the net difference. For example: if you are on day 15 of a monthly cycle at €30 and upgrade to a €60 plan, you pay approximately €15 (half the additional monthly cost). New limits take effect immediately after confirmation — no need to wait for the next renewal.

Yes. From the Billing page you'll find the option to downgrade to the Free plan. The downgrade is scheduled for the end of the current cycle and is not immediate.

All data and extracted records are preserved. The new Free plan limits apply from the start of the next cycle. Features or resources that exceed the new plan limits may no longer be available.

No. There is no pro-rata refund for unused days of the current billing cycle following a downgrade. Service remains available at the current plan until the end of the already-paid cycle.

If a Stripe subscription is active, it is automatically cancelled at end of period when the downgrade is confirmed. No further charges will be made for the next plan.

Once confirmed, the downgrade is scheduled and cannot be cancelled through the interface. If you need to keep the current plan, contact support before the end of the cycle.

To cancel your account contact support through the Contact page. Before cancellation you'll be asked to export your data if you wish.

Yes. From the Frontier page you can export records in CSV or Excel format. You can also request a full export through our support.

Data is retained for a period defined by our retention policies (configurable per plan) and then permanently deleted. For details see the Privacy Policy.

If enforcement is active for your plan, new AI processing is blocked (error 402) and IMAP ingestion is suspended until the next cycle starts. You can upgrade at any time to restore service.

Yes. When you use 80% of your monthly quota you receive a warning via email and a banner in the interface. You can manage notification preferences in Settings.

Each plan has limits on the number of documents that can be processed per day, week and month. Exceeding a rate limit returns a 429 error and temporarily blocks processing for the corresponding window (resets automatically).

Yes, through add-ons you can purchase extra tokens, documents or storage that add to your base plan limits. Add-ons are assigned from the administration panel.

Data is processed in accordance with GDPR. See our Privacy Policy for full details on collection, retention and data subject rights.

No. Documents and extracted data are not used to train third-party AI models. You can configure the AI provider used (including private on-premise provider) from the AI Config panel.

Attached files are stored on secure storage (MinIO S3-compatible). Retention is configurable per plan: attachments are automatically deleted when the retention period expires.

Yes. You have the right to request deletion of your personal data under GDPR. Contact us through the Contact page.

On first login a guided card appears with 4 steps: email verification, business profile completion (VAT number, SDI code etc.), billing plan selection and optional WhatsApp channel setup. You can complete steps in any order and resume at any time from the Onboarding page in the sidebar.

After completing your profile, go to Mailboxes and add your first IMAP mailbox. Kadeon will immediately start polling for new emails and show them in the Frontier section for review.

Yes. The wizard is optional: you can close the card at any time by clicking 'Do it later'. The completion status of each step remains visible and accessible. To reopen the wizard, visit the Onboarding page from the sidebar.

From the Users panel (reserved for tenant_admin) you can invite operators by email. Each operator receives an activation link, can log in with email+password and sees only your tenant's records.

Any provider that supports standard IMAP: Gmail (with app password), Outlook, Yahoo, PEC providers, company Exchange/IMAP servers and many others. You just need the host, port and IMAP credentials.

Yes. IMAP credentials are encrypted at the application level before being saved to the database and are never exposed in plain text via the API. Only internal polling processes can temporarily decrypt them to make the connection.

Polling happens every few minutes. You can force an instant manual check with the 'Check new emails' button in the Frontier section. Frequency may vary depending on the plan.

Yes. You can connect multiple mailboxes to the same account. The maximum number depends on the subscribed plan. Each mailbox can have routing rules and a dedicated outgoing SMTP for sending notifications.

Yes. When configuring a mailbox you can add SMTP parameters (host, port, username, password). This allows you to send emails — document notifications, replies — using that mailbox as the sender.

A document identity represents one of your company names or operating units. It defines the name, SDI code, trusted email addresses and routing rules. Kadeon uses identities to automatically match each incoming document to the right structure and extract data in the correct format.

The number depends on the plan. The Free plan includes a limited number of active identities; Subscription and Enterprise plans include unlimited or high-quota identities.

Not mandatory but strongly recommended if you receive electronic invoices. The SDI code allows Kadeon to correctly route XML documents to the right identity and associate them automatically without manual review.

It depends on the platform configuration. In self-service mode you can create identities autonomously and activate them immediately. In approval mode, the request is sent to the Kadeon team who validate it before activation.

Kadeon analyses the text and attachments of each email using an AI model. The AI identifies and structures key data (amounts, dates, tax codes, invoice numbers, references) and presents them to the operator for review in the Frontier section. You can customise extraction instructions per document identity.

By default Kadeon uses a local (on-premise) AI engine on your server. If you wish you can configure an external cloud provider (Anthropic Claude API) as fallback or primary provider from the AI Config panel. The choice can be changed at any time.

Only if you want to use the Claude model as an external AI provider. The key is personal, not included in any Kadeon plan, and must be created on console.anthropic.com. If you only use the local model no external key is needed.

Only if you have configured and enabled the Anthropic provider as fallback or primary provider. If the cloud fallback is disabled from the AI Config panel, all documents are processed exclusively on the local server with no external transmission.

The WhatsApp Business Bridge lets you send document routing notifications (e.g. receipt confirmation, document ready alert) via WhatsApp messages to your contacts or internal teams. Configure it from the WhatsApp panel in tenant settings.

You need a WhatsApp Business API token (Meta for Developers) and a Phone Number ID linked to your Meta Business account. The token is personal and is not included in any Kadeon plan. See Meta documentation for the registration process.

No. Kadeon uses WhatsApp exclusively for notification routing and does not store the content of WhatsApp conversations. Messages pass through the Meta platform subject to its terms of service.

WhatsApp Business Bridge activation is available as an add-on feature. Availability per plan is shown on the Pricing page.

From inside the dashboard go to Support. From there you can open a new ticket, send messages and follow all conversations with the team in real time. No need to leave the platform.

Yes. The Support section shows the current operator status: online (fast response expected), busy (working on other tickets), offline (reply when back). You can still send messages at any time: they are delivered as soon as the operator is available.

Each conversation has a subject and a series of messages. You can open multiple tickets for different topics. Tickets are closed by the support operator once resolved. If you still need help, you can reopen it by sending a new message.

Tickets are visible to all users with the tenant_admin role on your account. Operators on your tenant can view open tickets but cannot create new ones — that function is reserved for the tenant_admin.

Kadeon exposes an HTTP endpoint that accepts emails (body + base64 attachments) via POST. You can call this endpoint from your ERP, an RPA automation, or any script. The service processes the document with AI and brings it to the Inbox for review. Once approved, you can use the platform's native ERP integration to transfer the structured data.

Go to Settings and scroll to the API Keys — ERP Integration section. Click «New API key», give it a descriptive label (e.g. «Production ERP») and copy the key immediately — it will not be shown again. You can generate multiple keys (one per system) and revoke any you no longer need.

Add the HTTP header X-API-Key: mp_xxxxxxxx… to every request to the ingestion endpoint. The tenant is identified automatically from the key — no need to pass the tenant_id in the body. See the API documentation for the full payload format (email body, base64 attachments, optional fields).

The M2M ingestion endpoint is: POST /api/v1/ingest/email. The body is JSON with fields subject, from, body_text and an attachments array (each attachment has filename, mime and content_b64 — the file in base64). The response is 202 Accepted with details of the created record. See the API documentation for a full example.

No. Keys do not expire. You can however revoke them at any time from the API Keys section in Settings. A revoked key stops working immediately; it cannot be re-enabled, but you can generate a new one. For security we recommend creating a separate key for each system or integration and revoking any unused ones.

There is no preset limit on the number of keys. You can create as many as needed — one per system, environment (test/production), or automated script. Each key carries a descriptive label so you can identify it at a glance in the panel.