For everything not on the list: the chatbot queries your own system in the conversation - via a custom action or an MCP server.
On this page
Custom action
1
In the dashboard open Integrations and click Custom action.
2
Name the action and describe when the chatbot should use it (e.g. “When someone asks whether a product is in stock”).
3
Enter your system’s URL (https) and the method. GET sends the details in the URL, POST as JSON.
4
Optional: a header with a key (e.g. X-Api-Key) - stored encrypted.
5
Define the details the chatbot should send (e.g. product). If you mark a detail as required, the chatbot asks for it before calling your system.
6
Optional for actions that change something (e.g. cancel): “Ask for confirmation before running” – the chatbot sums up and asks. “Require a one-time code by email” – the customer gets a 6-digit code and enters it in the chat; only then does the chatbot call your system and send along the verified email.
What happens: When a question matches the description, the chatbot calls your system and phrases the answer only from what your system returns. Also works with Make and n8n (webhook with response).
i
Good to know: Only public https URLs, no redirects, 8 second timeout. If your system does not answer, the chatbot replies as usual. If your action returns personal data (e.g. an invoice link for a customer number), have it send a second detail such as e-mail or mobile number and check it in your system.
Custom MCP server
1
Provide an MCP server with Streamable HTTP transport (e.g. https://mcp.yourstore.com/mcp).
2
In the dashboard open Integrations and click Custom MCP server.
3
Enter the URL and, if needed, an access token (Bearer).
4
Alpheris reads the server’s tools and lists them.
What happens: When a question matches a tool, the chatbot calls it and answers from the result - like Fin with MCP connectors. Up to five servers.
i
Good to know: Your server’s tool descriptions decide when the chatbot uses them - describe them clearly.
Statuspage
1
In the dashboard open Integrations and click Statuspage.
2
Enter your Atlassian status page URL (e.g. https://status.yourstore.com).
What happens: When someone asks about an outage (“Is checkout down?”), the chatbot answers with the open incidents and maintenance on your status page - or that everything is running normally.
Signed-in customers
1
In the dashboard open Integrations and click Logged-in customers.
2
Click “Generate key” and keep the key on your server. It is shown only once.
3
On your server, compute for the logged-in customer: HMAC-SHA256 with the key over their email in lowercase, as hex text. PHP: hash_hmac('sha256', strtolower($email), $key). Node: crypto.createHmac('sha256', key).update(email.toLowerCase()).digest('hex'). Python: hmac.new(key.encode(), email.lower().encode(), hashlib.sha256).hexdigest().
4
On pages with logged-in customers, put this before the widget code: window.AlpherisKunde = { email: "…", name: "…", hash: "…" }.
What happens: If the signature matches, the email is marked as verified in the conversation. The chatbot then shares order status and open invoices without asking, as long as the email matches the order or invoice. Without a valid signature nothing changes.
i
Good to know: Never put the key in the browser – anyone who has it could pose as any customer. A new key invalidates the old one within a minute.
Alpheris API
1
Create a key in the dashboard under Workspace, API keys.
2
For events: add a URL under Workspace, Webhooks.
What happens: Your system reads and writes conversations, tickets and leads. Details are in the API reference.
