Settings

System configuration, database connections, theme customization, and agent registry — all managed from a single page.

Why Use Settings?

The Settings page is your control center for connecting AXIS to your infrastructure. Instead of editing YAML files manually for every change, the Settings page gives you a live view of the current configuration and lets you verify connections at a glance.

Quick Start

Here is the recommended workflow when setting up AXIS for the first time:

Backend Connection

The first card on the Settings page shows the API endpoint URL and a live connection status indicator. AXIS checks the backend /health endpoint on page load.

Database Configuration

AXIS supports four independent database connections, each configured through a dedicated YAML file. All databases are optional — the platform works offline with CSV uploads alone.

Connection Fields

Each database config shares the same base connection fields:

Example YAML

Here is a minimal eval_db.yaml configuration:

eval_db:
  enabled: true
  auto_load: false
  host: db.example.com
  port: 5432
  database: evaluations
  username: axis_reader
  password: s3cret
  ssl_mode: require
  dataset_query: "SELECT * FROM evaluation_dataset"
  results_query: "SELECT * FROM evaluation_results"
  query_timeout: 60
  row_limit: 10000

Sync Engine (DuckDB)

AXIS uses an embedded DuckDB store as a local analytics cache. The sync engine pulls data from configured PostgreSQL databases and stores it locally for fast querying.

To trigger a manual sync, call the API endpoint:

POST /api/store/sync

Thresholds

AXIS uses numeric thresholds to determine pass/fail status and color-code metrics across dashboards. These are defined in the backend config.py and the monitoring YAML config.

Global Thresholds

Monitoring Thresholds

The monitoring database config supports per-source threshold overrides:

monitoring_db:
  thresholds:
    default:
      good: 0.7
      pass: 0.5
    per_source:
      alpha_bot:
        good: 0.8
        pass: 0.6
      beta_bot:
        good: 0.75
        pass: 0.55

Per-source thresholds let you set stricter standards for critical agents while keeping defaults for the rest.

Feature Flags

Feature flags control which capabilities are available in the frontend. They are served by the /api/config/features endpoint and configured in the backend YAML files.

Theme & Branding

AXIS is white-label ready. All branding strings and colors are configured through theme.yaml or environment variables — never hardcoded in the frontend.

Color Palettes

Two built-in palettes are available, and you can define custom palettes in the YAML:

Branding Fields

These text fields appear throughout the application — in the sidebar, hero section, reports, and page headers:

# theme.yaml
theme:
  active: professional_blue
  branding:
    app_name: "MyCompany AI"
    tagline: "Internal Evaluation Platform"
    subtitle: "Quality Assurance Suite"
    description: "AI evaluation and monitoring for MyCompany"
    report_footer: "Generated by MyCompany AI Platform"

Agent Configuration

The agent registry tells AXIS about the AI agents in your system. Agents appear in the SourceSelector dropdown and production dashboards. Configured in agents.yaml.

# agents.yaml
agents:
  - name: alpha_bot
    label: Alpha Bot
    role: Customer Support
    avatar: /agents/alpha_bot.png
    description: Handles inbound customer requests
    active: true
  - name: beta_bot
    label: Beta Bot
    role: Order Processing
    avatar: /agents/beta_bot.png
    description: Manages order workflows end to end
    active: true

Config Files Reference

All configuration files live in backend/config/. Only .example templates are version-controlled. Run make setup to create your local copies.

Configuration Load Order

For most config files, AXIS follows this precedence (highest wins):

1. YAML config file — if present in backend/config/
2. Environment variables — override specific YAML values
3. Hardcoded defaults — in config.py dataclasses

Next Steps