Workjet
Back to Workjet

Gateway Admin

The gateway admin section in the Portal lets admins and owners configure how AI requests are routed, how sensitive data is protected, and which models are available to the team. Changes made here apply immediately to all users in the organization.

Configuring Routing Rules

The routing configuration determines which AI provider and model handles each request tier. See Model Routing for a detailed explanation of tiers, providers, and fallback.

  1. Navigate to Gateway in the Portal sidebar
  2. Select the Routing tab
  3. For each tier (Standard, Premium, Enterprise):
    • Click Add Model
    • Select the provider (Anthropic, OpenAI, Google, Ollama)
    • Choose the model
    • Set the priority (lower number = higher priority)
    • Set max input and output token limits
  4. Click Save Configuration

Tip: Add at least two models per tier for redundancy. If the primary provider has an outage, the gateway automatically routes to the fallback model.

Managing DLP Policies

The DLP section lets you control which data patterns are scanned and what happens when sensitive data is detected. See DLP Policies for full details.

  1. Navigate to Gateway in the Portal sidebar
  2. Select the DLP tab
  3. Toggle the global DLP switch to enable or disable scanning
  4. For each built-in pattern, choose the action: Redact, Block, Warn, or Disabled
  5. To add a custom pattern:
    • Click Add Custom Pattern
    • Enter a name, regex pattern, and action
  6. Click Save Policies

Viewing Current Configuration

The gateway admin page displays the current active configuration at the top, including:

  • Active models: List of all configured models with their tier and priority
  • DLP status: Whether DLP is enabled and how many patterns are active
  • Rate limits: Current per-user rate limit thresholds
  • Budget status: Current spend against the monthly cap

This summary provides a quick health check of your gateway configuration without needing to open each settings tab.

Provider API Keys

The gateway needs API keys for each AI provider you want to use. These are entered in the gateway admin under the Providers tab:

Provider Key Type Where to Get It
Anthropic API Key console.anthropic.com
OpenAI API Key platform.openai.com/api-keys
Google API Key aistudio.google.com/apikey
Ollama Base URL Your local Ollama server (e.g., http://localhost:11434)

Security: Provider API keys are encrypted at rest and never exposed in the Portal UI after initial entry. Only the owner can view or rotate provider keys.

Rate Limiting

Configure per-user rate limits from the gateway admin:

  • Window: The time window for counting requests (e.g., 1 minute, 1 hour)
  • Threshold: Maximum number of requests per window

When a user exceeds the limit, they receive a 429 Too Many Requests response with standard X-RateLimit-* headers indicating when they can retry.

Next Steps