Quickstart¶

Get Orion running and trigger your first content pipeline in under 5 minutes. This guide walks you through the complete flow — from starting the platform to generating and approving content.

Start the Platform¶

cp .env.example .env
docker compose -f deploy/docker-compose.yml up -d

Wait for all services to become healthy:

# Watch container health status
docker compose -f deploy/docker-compose.yml ps

# Or check through the CLI
cd cli && uv run orion system health

Authenticate¶

You need a JWT token to interact with the API. The default development credentials are admin@orion.local / orion_dev.

orion auth login
# Email: admin@orion.local
# Password: orion_dev

orion auth whoami

TOKEN=$(curl -s http://localhost:8000/api/v1/auth/login \
  -H "Content-Type: application/json" \
  -d '{"email": "admin@orion.local", "password": "orion_dev"}' \
  | jq -r '.access_token')

echo $TOKEN

import httpx

resp = httpx.post(
    "http://localhost:8000/api/v1/auth/login",
    json={"email": "admin@orion.local", "password": "orion_dev"},
)
token = resp.json()["access_token"]
print(f"Token: {token[:20]}...")

Trigger a Trend Scan¶

Kick off the Scout service to detect trending topics from external sources.

# Trigger a trend scan
orion scout trigger

curl -X POST http://localhost:8000/api/v1/scout/api/v1/trends/scan \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"region": "US", "limit": 10}'

resp = httpx.post(
    "http://localhost:8000/api/v1/scout/api/v1/trends/scan",
    headers={"Authorization": f"Bearer {token}"},
    json={"region": "US", "limit": 10},
)
print(resp.json())
# {"message": "Scan complete", "trends_found": 10, "trends_saved": 7}

View Detected Trends¶

After a scan completes, Scout publishes orion.trend.detected events to Redis. You can query the results immediately.

# List detected trends
orion scout list-trends

# List trends in JSON format
orion scout list-trends --format json

curl http://localhost:8000/api/v1/scout/api/v1/trends \
  -H "Authorization: Bearer $TOKEN"

Generate Content¶

When Scout detects a trend, the Director service automatically picks it up via Redis pub/sub and starts a LangGraph pipeline. You can also trigger content generation manually.

# List all content items
orion content list

# List content in JSON format
orion content list --format json

curl -X POST http://localhost:8000/api/v1/director/api/v1/content/generate \
  -H "Authorization: Bearer $TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "trend_id": "<trend-uuid>",
    "trend_topic": "AI agents in production",
    "niche": "technology",
    "target_platform": "youtube_shorts"
  }'

Review, Approve, and Publish¶

Content goes through a human-in-the-loop review stage before publishing.

# List all content items
orion content list

# View full details of a content item (script, assets, metadata)
orion content view <content-id>

# Approve content for publishing
orion content approve <content-id>

# Reject content
orion content reject <content-id>

# Request regeneration
orion content regenerate <content-id>

# Check final publish status
orion content view <content-id>

What Happens Behind the Scenes¶

When you trigger a scan, the following event-driven pipeline executes automatically:

sequenceDiagram
    participant User
    participant Gateway
    participant Scout
    participant Redis
    participant Director
    participant Media
    participant Editor

    User->>Gateway: POST /api/v1/scout/api/v1/trends/scan
    Gateway->>Scout: Proxy request
    Scout->>Scout: Poll external sources
    Scout->>Redis: Publish orion.trend.detected
    Redis->>Director: Deliver event
    Director->>Director: LangGraph pipeline (strategist -> creator)
    Director->>Redis: Publish orion.content.created
    Redis->>Media: Deliver event
    Media->>Media: Generate images (ComfyUI/Fal.ai)
    Media->>Redis: Publish orion.media.generated
    Redis->>Editor: Deliver event
    Editor->>Editor: Render video (TTS + stitch + subtitles)

Each service communicates exclusively through Redis pub/sub — there are no direct HTTP calls between Python services. This decoupled architecture means any service can be restarted or scaled independently without affecting the pipeline.

Next Steps¶

• Configuration — Tune environment variables, rate limits, and AI providers
• CLI Reference — Full list of all CLI commands and flags
• API Endpoints — REST API documentation for programmatic access
• LangGraph Pipeline — Understand the content creation graph
• Docker Deployment — Production deployment and scaling