Skip to main content

​
Overview

Environment variables allow you to:
  • Store API keys and secrets securely
  • Configure runtime behavior without code changes
  • Use different settings across environments

​
Setting Variables

​
Via CLI

crewship env set OPENAI_API_KEY=sk-proj-...

​
Via Console

  1. Open the Crewship Console
  2. Navigate to your project
  3. Go to Settings → Environment Variables
  4. Click Add Variable

​
Common Variables

​
LLM API Keys

# OpenAI
crewship env set OPENAI_API_KEY=sk-proj-...

# Anthropic
crewship env set ANTHROPIC_API_KEY=sk-ant-...

# Google AI
crewship env set GOOGLE_API_KEY=...

# Azure OpenAI
crewship env set AZURE_OPENAI_API_KEY=...
crewship env set AZURE_OPENAI_ENDPOINT=https://...

​
Tool API Keys

# Web search (Serper)
crewship env set SERPER_API_KEY=...

# Browserless
crewship env set BROWSERLESS_API_KEY=...

# Firecrawl
crewship env set FIRECRAWL_API_KEY=...

​
Custom Configuration

# Logging level
crewship env set LOG_LEVEL=debug

# Feature flags
crewship env set ENABLE_CACHING=true

# Custom settings
crewship env set MAX_RETRIES=3

​
Accessing in Code

Variables are available as standard environment variables: 
import os

# Direct access
api_key = os.environ["OPENAI_API_KEY"]

# With default
log_level = os.environ.get("LOG_LEVEL", "info")

# Check if set
if os.environ.get("ENABLE_CACHING"):
    enable_cache()

​
CrewAI Integration

CrewAI automatically uses standard environment variables: 
from crewai import Agent, LLM

# Uses OPENAI_API_KEY automatically
agent = Agent(
    role="Researcher",
    llm=LLM(model="gpt-4o")
)

# Anthropic (uses ANTHROPIC_API_KEY)
agent = Agent(
    role="Writer",
    llm=LLM(model="claude-3-5-sonnet-20241022")
)

​
Reserved Variables

Crewship sets these automatically. They’re read-only:
VariableDescription
CREWSHIP_RUN_IDCurrent run ID (run_abc123)
CREWSHIP_DEPLOYMENT_IDDeployment ID (dep_xyz789)
CREWSHIP_PROJECTProject name
CREWSHIP_ENVIRONMENTproduction or preview
Use them for logging and debugging: 
import os

run_id = os.environ.get("CREWSHIP_RUN_ID")
print(f"Starting run {run_id}")

​
Security

​
Encryption

All environment variables are:
  • Encrypted at rest using AES-256
  • Encrypted in transit via TLS
  • Never logged in build or run output
  • Access controlled by project permissions

​
Best Practices

Add .env to .gitignore. Use crewship env set instead.
Create API keys specific to Crewship. Easier to rotate and audit.
Don’t use fallback values for required secrets:
# ❌ Bad - silently fails
api_key = os.environ.get("API_KEY", "")

# ✅ Good - fails immediately
api_key = os.environ["API_KEY"]
List required variables in your README:
## Environment Variables

- `OPENAI_API_KEY` - Required. OpenAI API key
- `SERPER_API_KEY` - Required for web search
- `LOG_LEVEL` - Optional. Default: info

​
Loading from .env

For local development, use a .env file:
.env
OPENAI_API_KEY=sk-proj-...
SERPER_API_KEY=...
LOG_LEVEL=debug
Then load with python-dotenv: 
from dotenv import load_dotenv
load_dotenv()
The .env file is not used in Crewship deployments. Always set production variables via crewship env set.

​
Variable Scope

Variables are scoped to projects, not deployments: 
Project: my-crew
├── OPENAI_API_KEY = sk-...
├── SERPER_API_KEY = ...
│
├── Deployment: dep_abc (uses these vars)
├── Deployment: dep_xyz (uses these vars)
└── All runs use these vars

​
Updating Variables

Changes take effect immediately for new runs: 
# Update a variable
crewship env set OPENAI_API_KEY=sk-new-key-...

# Next run uses new value (no redeploy needed)
crewship invoke --input '{"topic": "test"}'

​
Environment Separation

Use separate projects for different environments: 
# Production
crewship env set OPENAI_API_KEY=sk-prod-... --project my-crew
crewship env set LOG_LEVEL=warning --project my-crew

# Staging
crewship env set OPENAI_API_KEY=sk-test-... --project my-crew-staging
crewship env set LOG_LEVEL=debug --project my-crew-staging

CLI env Command

Full CLI reference

Configuration

Project configuration