Skip to main content

​
What are Projects?

Projects help you organize your LLM integrations within an organization. Each project contains:
  • API Keys: Credentials for making gateway requests
  • Provider Keys: Optional upstream provider API keys
  • Usage Logs: Request history and analytics
  • Caching Settings: Response caching configuration
  • Mode Configuration: How requests are processed (credits vs. API keys)
Each organization can have up to 10 projects. Contact support to increase this limit.

​
Use Cases

Environment Separation

Create separate projects for development, staging, and production environments.

Application Isolation

Isolate different applications or services with dedicated projects.

Team Segmentation

Give different teams their own projects with separate billing and usage tracking.

Client Management

Manage multiple client implementations with separate projects and API keys.

​
Creating a Project

​
Via Dashboard

1

Navigate to Projects

From your organization dashboard, click Projects in the sidebar.
2

Create New Project

Click the New Project button in the top right.
3

Configure Settings

Enter your project details:
  • Name: A descriptive name (e.g., “Production API”, “Mobile App”)
  • Mode: Choose how requests are processed
  • Caching: Enable or disable response caching
4

Create API Key

After creating the project, generate your first API key to start making requests.

​
Via API

Create a project programmatically: 
curl -X POST https://api.llmgateway.io/api/v1/projects \
  -H "Authorization: Bearer YOUR_SESSION_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Production API",
    "organizationId": "org_123abc",
    "mode": "hybrid",
    "cachingEnabled": true,
    "cacheDurationSeconds": 300
  }'
Response: 
{
  "project": {
    "id": "proj_456def",
    "name": "Production API",
    "organizationId": "org_123abc",
    "mode": "hybrid",
    "cachingEnabled": true,
    "cacheDurationSeconds": 300,
    "status": "active",
    "createdAt": "2024-03-15T10:30:00Z",
    "updatedAt": "2024-03-15T10:30:00Z"
  }
}

​
Project Modes

Projects support three operating modes that determine how LLM requests are processed: 
{
  "mode": "hybrid"
}
Requests automatically use:
  1. Your provider API keys if configured (no gateway fees)
  2. Gateway credits as fallback (with gateway fees)
Hybrid mode gives you the flexibility to use your own API keys when available while maintaining uninterrupted service through credits.

​
API Keys Mode

{
  "mode": "api-keys"
}
All requests require provider API keys configured in your organization. No credits are used. Use when:
  • You want direct provider billing
  • You need maximum cost control
  • You have enterprise agreements with providers

​
Credits Mode

{
  "mode": "credits"
}
All requests use gateway credits. Provider API keys are ignored. Use when:
  • You want simplified billing
  • You need predictable per-token pricing
  • You’re prototyping without provider accounts

​
Response Caching

Enable caching to reduce costs and improve response times for repeated queries.

​
How It Works

  1. Gateway caches responses based on request parameters
  2. Identical requests return cached responses instantly
  3. Cache entries expire after the configured duration
  4. Cached responses don’t count toward usage costs

​
Configuration

Enable via Dashboard: Go to Project Settings → Caching and toggle the setting. Enable via API: 
curl -X PATCH https://api.llmgateway.io/api/v1/projects/proj_456def \
  -H "Authorization: Bearer YOUR_SESSION_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "cachingEnabled": true,
    "cacheDurationSeconds": 300
  }'

​
Cache Duration

Set how long responses are cached (10 seconds to 1 year):
  • 10-60 seconds: Real-time data with light caching
  • 5-15 minutes: General queries with moderate caching
  • 1-24 hours: Static content or documentation queries
  • Days/weeks: Unchanging content like historical data
Caching is based on the complete request body. Even small changes in parameters result in cache misses.

​
Best Practices

  • User-facing chatbots with common questions
  • Documentation or FAQ systems
  • Content generation with templates
  • Repeated data analysis queries
  • Real-time data queries
  • Personalized responses
  • Time-sensitive information
  • Unique user inputs

​
Managing Projects

​
Update Project Settings

Modify project configuration: 
curl -X PATCH https://api.llmgateway.io/api/v1/projects/proj_456def \
  -H "Authorization: Bearer YOUR_SESSION_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Production API v2",
    "mode": "api-keys",
    "cachingEnabled": false
  }'

​
Get Project Details

Retrieve project information: 
curl https://api.llmgateway.io/api/v1/projects/proj_456def \
  -H "Authorization: Bearer YOUR_SESSION_TOKEN"

​
Delete a Project

Permanently delete a project (requires owner role): 
curl -X DELETE https://api.llmgateway.io/api/v1/projects/proj_456def \
  -H "Authorization: Bearer YOUR_SESSION_TOKEN"
Deleting a project will:
  • Invalidate all API keys
  • Remove all usage logs
  • Stop all active requests
This action cannot be undone.

​
API Keys

API keys authenticate your requests to the gateway. Each project can have multiple keys.

​
Creating API Keys

1

Navigate to Project

Select your project from the dashboard.
2

Go to API Keys Tab

Click API Keys in the project menu.
3

Generate New Key

Click Create API Key and provide:
  • Description: What this key is used for
  • Usage Limit (optional): Maximum spend allowed
4

Save Your Key

Copy and securely store your key. It won’t be shown again.

​
Usage Limits

Set per-key spending limits to control costs: 
curl -X POST https://api.llmgateway.io/api/v1/api-keys \
  -H "Authorization: Bearer YOUR_SESSION_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "projectId": "proj_456def",
    "description": "Production API Key",
    "usageLimit": 100.00
  }'
Keys are automatically disabled when they exceed their usage limit.

​
IAM Rules

Restrict what models and providers an API key can access. See the API Keys guide for details.

​
Best Practices

  • Use one project per environment (dev, staging, prod)
  • Name projects clearly: “Mobile App - Production”
  • Document project purposes in descriptions
  • Review and archive unused projects regularly
  • Rotate API keys periodically
  • Use different keys for different services
  • Set usage limits on all keys
  • Delete unused keys immediately
  • Never commit keys to version control
  • Enable caching for repeated queries
  • Set usage limits on API keys
  • Use hybrid mode to leverage your provider keys
  • Monitor usage regularly in the dashboard
  • Set up billing alerts
  • Review logs for errors and anomalies
  • Track token usage by model
  • Monitor cache hit rates
  • Set up alerts for spending thresholds
  • Analyze performance metrics

​
Next Steps

API Keys

Learn about API key management and IAM rules.

Provider Keys

Configure upstream provider API keys.

Usage Analytics

Monitor and analyze your gateway usage.

Playground

Test models in the interactive playground.

Build docs developers (and LLMs) love

Get started for freeTalk to us