Skip to main content

AI Applications

Deploy and manage containerized applications with integrated access to databases, AI models, and workflows.

What are AI Applications?​

AI Applications on the Strongly platform are containerized workloads that can access platform services seamlessly. Whether you're deploying a React SPA, Node.js API, Flask backend, R Shiny dashboard, or custom application, the platform provides:

  • Automated builds from your code archives, GitHub repos, or volume folders (the code you sync from a workspace)
  • Integrated service connections (databases, AI models, workflows, ML models)
  • Auto-scaling on CPU and memory use, between the instance counts you set
  • Health monitoring and resource metrics
  • GPU deployment support for compute-intensive workloads

Supported Application Types​

TypeDefault PortUse Case
react3000React/Vue SPA with client-side routing
nodejs3000Node.js APIs, Express, NestJS
flask5000Python Flask/FastAPI applications
rshiny3838R Shiny dashboards and interactive apps
static80Static HTML/CSS/JS sites
fullstack3000, 8000Monorepo with frontend + backend
mcp_server8080MCP protocol tool servers
custom8080Any other application type

Key Features​

Service Integration​

Connect your application to platform resources automatically:

  • Add-ons: Managed MongoDB, PostgreSQL, MySQL, Redis, RabbitMQ, Neo4j, Milvus, Greenplum, SurrealDB instances
  • Data Sources: External databases, S3, Snowflake, BigQuery, and many more
  • AI Models: OpenAI, Anthropic, or self-hosted LLM models via AI Gateway
  • Workflows: Invoke Strongly workflows via REST API
  • ML Models: Traditional and AutoML models from the model registry

All connections are injected via the STRONGLY_SERVICES environment variable - no manual configuration needed.

Auto-Scaling​

Enable intelligent auto-scaling based on CPU and memory thresholds:

  • Scale from 1 to 50 instances automatically
  • Custom thresholds and polling intervals
  • Configurable cooldown periods to prevent flapping

Environments​

Every app deploys into an Environment -- a named compute configuration (CPU, memory, and optionally GPU) managed by your administrators under Operations > Environments. At deploy time you pick an environment from the dropdown, or choose Custom Configuration to set CPU, memory, disk, and GPU directly. A Use Spot Instances option can reduce compute cost (up to 70%) for workloads that tolerate interruption.

Quick Start​

  1. Prepare your code: Include Dockerfile and strongly.manifest.yaml
  2. Create archive: Package as .zip, .tar, .tar.gz, or .tgz
  3. Deploy: Navigate to Apps and deploy your application
  4. Configure: Set resources, environment variables, and service connections
  5. Monitor: View logs, metrics, and health status

Deployment Sources​

Applications can be deployed from multiple sources:

  • File upload: Upload a .zip, .tar, .tar.gz, or .tgz archive
  • GitHub: Connect a GitHub repository directly
  • Project bundle: Deploy from platform project bundles

Next Steps​