Documentation Index
Fetch the complete documentation index at: https://docs.elementum.io/llms.txt
Use this file to discover all available pages before exploring further.
Required Permission:
Overview
Agent Gateways provide a single entrypoint to multiple agents across your organization. Instead of users navigating to individual agents in different apps, a gateway consolidates them behind one interface. Users interact with the gateway, which routes their requests to the appropriate agent based on context. When enabled as the front door, the gateway becomes the primary assistant in the Action Center on the Home page — giving every user in the organization a consistent starting point for AI-powered assistance.Agent Gateways require at least one configured AI Provider with an available LLM. See AI Services to set up providers and models.
Create a Gateway
- Open
Org Settings → Agent Gateways in the Platform section.
- Click New Gateway.
- Enter a Name and Description for the gateway.
- Select an AI Provider.
- Click Create Gateway.
Configure the Gateway
Click on a gateway to open its configuration. Use the toggle at the top of the page to make the gateway available to users.Agents
Select the Agents tab to manage which agents are accessible through the gateway.- Add agents from across your organization to make them available through this gateway
- Toggle the Platform Assistant on if you want to include Elementum’s built-in assistant alongside your custom agents. See Platform Assistant (Elementum Edge) to configure it first.
Platform Assistant (Elementum Edge)
Elementum Edge is the built-in platform assistant for an Agent Gateway. It provides general AI help to gateway users when no specific process agent is the right fit. The assistant tailors its responses to each user based on their Personalization settings in the gateway. To enable and configure the platform assistant:- Click the
Settings icon in the top-right corner of the gateway.
- In the Platform Assistant tile, click Configure Platform Assistant.
- Toggle on Elementum Edge.
- Configure the agent:
- Name — Display name for the assistant. Defaults to Elementum Edge.
- Model — Select the LLM that powers the assistant.
- Click Save.
- On the Gateway Settings page, toggle on the platform assistant. If you changed the name, the new name appears here in place of Elementum Edge.
Rules
Select the Rules tab to define how the gateway routes incoming requests to specific agents. To create a rule:- Click + Add Rule.
- Choose your rule type:
- Deterministic — Add a name, description, and conditions to filter the rule. Use for structured requests that are easy to filter (for example, channel, user, or department).
- Probabilistic — Add a name, description, and conditional description (AI prompt) for the LLM to use, then choose the desired routing from the dropdown. Use for natural language requests that are evaluated by AI to select the correct route.
- Click Update Rule.
Deterministic rules are evaluated before probabilistic rules during a chat with a user.
Settings
Select the Settings tab to update the gateway’s core configuration.- Name and Description — Update the gateway’s display name and description.
- AI Provider — Change the AI provider used by the gateway.
- Front Door — Enable this toggle to set the gateway as the primary assistant in the Action Center on the Home page. This makes the gateway the default AI interaction point for all users in your organization.
Test
Select the Test tab to verify the gateway is working as intended before enabling it for your organization. Send messages and confirm that requests are routed to the correct agents.Personalization
Personalization is available to all users of an Agent Gateway, not just administrators. Each user configures their own personalization independently, and it only affects their own experience with the gateway.
- From the gateway Home page, click the
Settings icon in the bottom-left corner.
- Select Personalization.
- Configure the following:
- User Info — Details about your role, team, or other identifying information.
- Goals — Personal goals and responsibilities the assistant should consider when responding.
- Relevant Apps — Apps that contain the data and processes most relevant to your work.
Related Features
AI Services
Configure providers and models that power your gateway
Building Agents
Create and configure the agents available through your gateway
Agent tools, deployment, and integrations
Multi-agent coordination and the A2A protocol for agent-to-agent communication
Agent Skills
Define reusable skill modules that gateway agents can use