System Settings

Configure LLM providers, rate limits, notifications, and security. Changes require service restart.

Configuration changes require restart

Most settings are loaded at service startup. After saving, restart the affected service from the Services page.

LLM / AI Configuration

Manage API keys and model selection for the multi-model cascade.

Claude API Key

Tier 2 -- Legal reasoning and fact-checking. Enable prompt caching for 80% cost reduction.

Gemini API Key

Tier 1 -- Routing, classification, language detection.

Claude Model

Used for legal document analysis and verification.

Gemini Model

Used for routing and classification tasks.

Embedding Model

Used for legal document vector embeddings.

Rate Limiting

Per-tier rate limits at the API gateway. Token bucket algorithm.

Free (Citizens)

req/min

Default tier for anonymous citizens.

Journalist

req/min

Verified journalists (free account).

Standard API

req/min

$99/month API consumers.

Enterprise API

req/min

$499/month enterprise tier.

Government

req/min

Government dashboards (contract).

Notification Channels

Email SMTP, SMS via Africa's Talking, and WhatsApp Business API.

SMTP Host

SMTP Port

Africa's Talking API Key

Used for USSD and SMS channels.

WhatsApp Cloud API Token

Meta Business Suite token for WhatsApp Bot.

Security & Authentication

JWT configuration, token lifetimes, and CORS origins.

JWT Access Token Expiry

hours

RS256/HS256 signed. Recommended: 24h.

Refresh Token Expiry

days

Stored in Redis with rotation.

CORS Allowed Origins

Comma-separated list of allowed origins.