What you can configure
Everything a tenant offers its users is set in Platform Admin under Tenants → \. The tenant page shows one card per area; the pencil on a card opens that area's editor, with a Form tab and a JSON tab for the same settings.
Creating a new tenant
This guide walks you through creating a new tenant using the Zeta Alpha Platform UI. The tenant creation process sets up a complete, ready-to-use environment including authentication, storage, a default index, and essential content sources.
Creating custom index
This guide walks you through creating a custom index tailored to your specific document structure and search requirements. By defining custom fields, search profiles, and display settings, you can optimize how documents are indexed, searched, and presented to users.
Adding a custom field to a default index
This guide walks you through converting a default index to a custom one in order to add new custom fields to an existing default index that was created after the creation of the tenant. In this way, you can extend the default index fields with new fields that may be needed to represent updates on your specific document structure and search requirements.
Generating API keys
API keys allow users and applications to authenticate programmatically with the Zeta Alpha platform without using passwords. This guide will walk you through the steps to create users and generate API keys for them.
Creating an analytics export template
An analytics export template defines a usage report for a tenant. Each run of a template (a
Configuring search filters
Search filters are the widgets shown in the search sidebar of the Zeta Alpha Navigator that let users narrow results by document metadata — for example, by document type, language, source, or tag. This guide walks you through adding a filter to an index, choosing the right widget, and configuring each widget type.
Managing LLM models
An LLM configuration is a named connection to one model at one vendor. Agents, dictation and image generation refer to it by name, and rules grant it by name (model:). Add one per model you want to offer.
Storing secrets
Store each API key, client secret or token once as a named tenant secret, and let LLM, tracing, agent and MCP server configurations refer to it by name. Rotating the key is then one edit.
Creating and editing agents
An agent is a model, a set of capabilities (tools, skills, memory, MCP servers) and the people it is shared with. Its display configurations decide how it appears in the Navigator. Create one per job your users need done, for example a research assistant or a meeting summarizer.
Connecting MCP servers
An MCP server gives agents tools from another system: Outlook, Teams, GitHub, HubSpot, a ticket system. Add the server to the tenant once, then mount it on each agent that should use it. Users sign in to it themselves, so the agent acts with each user's own rights.
Controlling who can use what
Rules decide who can use which agents, models, indexes, tools sources, MCP servers and tools, and whether a tool asks first and how it loads.
Requiring approval for actions
Make an agent ask the user before it runs a tool that sends, writes or deletes. The user sees an approval card and approves or declines the call. Set it per agent, or for every agent and some users with a setting rule.
Customizing the chat
Choose which agents users find where in the Navigator, how answers look, and which chat features are on. Agent placement and display are under the Client Settings card; the chat features are under the Features card.
Branding the Navigator
Give the tenant's Navigator your name, logo and colors, choose the page it opens on, and arrange the search page's widgets. All of it is under the Client Settings card: open Tenants → \ and click its pencil.
Tenant-wide agent settings
The Features card decides which gated capabilities the agent editor offers for every agent of the tenant. Open Tenants → \ and click the pencil on the Features card.
Exposing the tenant over MCP
Let Claude Code, Cursor, VS Code, Copilot Studio or another MCP client search and use the tenant's knowledge. You publish a bundle of tools as an MCP server setup; each person connects with their own sign-in or API key, and document access rights apply as in the Navigator.