Skip to main content
Projects are logical containers that organize traces, datasets, policies, and team members. They help you separate environments, applications, and teams.
Think of projects like repositories in GitHub—they’re the top-level organizational unit that keeps everything organized and separate.

​
What Is a Project?

A project groups together:
All execution records for this application or environment
Vector databases and knowledge bases connected to this project
Governance rules that apply to data usage
Who has access to view and modify this project

​
Why Use Projects?

Keep production data separate from staging and development:

Production

Live user traffic, real data

Staging

Pre-release testing, safe experimentation

Development

Local testing, rapid iteration
Each environment gets its own project, preventing cross-contamination.

​
Project Structure

Each project has:

​
Basic Information

  • Name - Human-readable identifier (“Production Customer Support”)
  • ID - Unique identifier used in API calls
  • Description - What this project is for
  • Created date - When it was set up

​
Connected Resources

  • Datasets - Which vector databases are linked

​
Metrics

  • Total traces - How many traces have been captured
  • Active datasets - How many data sources are connected
  • Usage - Token counts, costs, trace volume over time

​
Default Project

When you first set up your AI platform, a Default Project is typically automatically created. You can:
  • Use it immediately for quick testing
  • Rename it to match your main application
  • Create additional projects and delete the default

​
Creating a Project

1

Navigate to Projects

Go to the Projects page in your dashboard
2

Create new project

Click New Project button
3

Enter details

Provide a name and description for your project
4

Save

Click Create to finalize

​
Using a Project

​
Send Traces to a Project

When instrumenting your application, specify the project ID in your observability connector configuration: 
# Example instrumentation with project ID
connector = YourObservabilityConnector(
    base_url="https://your-platform.example",
    api_key="your-api-key-here",
    project_id="proj_abc123",  # Your project ID
)
connector.init()
All traces from this application will now appear in that project. Learn more about connecting your AI system →

​
Project Dashboard

Each project has a dashboard showing:

​
Trace Activity

  • Recent traces - Latest executions with status, cost, duration
  • Trace volume over time - Chart showing activity trends
  • Error rate - Percentage of failed traces
  • Average latency - Response time trends

​
Cost Tracking

  • Total spend - Cumulative cost for this project
  • Cost over time - Spending trends
  • Cost by model - Which models cost the most
  • Most expensive traces - Outliers to investigate

​
Data Usage

  • Active datasets - Number of connected data sources
  • Documents retrieved - Total retrieval count
  • Usage by dataset - Which datasets are used most
  • Unused documents - Percentage of docs never retrieved

​
Collections

  • Active collections - Curated trace groups for review
  • Shared collections - What’s been shared with stakeholders
  • Recent activity - Latest collection updates

​
Managing Projects

​
Rename a Project

Update the name or description:
  1. Open project settings
  2. Edit name/description
  3. Save changes

​
Delete a Project

Permanently remove a project and all its data:
  1. Open project settings
  2. Click Delete Project
  3. Type project name to confirm
  4. Click Delete Permanently
Warning: This deletes all traces, datasets, and collections. This cannot be undone.

​
Best Practices

Help team members understand what each project is:
❌ "Project 1"
❌ "Test"
❌ "My Project"
Create and configure projects before sending traces. This ensures traces go to the right place from the start.
Check the dashboard weekly:
Are error rates increasing?
Is cost trending up unexpectedly?
Are datasets being used effectively?

​
Common Workflows

1

Create a new project

Set up your project in the dashboard or via API
2

Configure project settings

Add a descriptive name and description
3

Instrument your application

Add the project ID to your application configuration
4

Deploy and monitor

Deploy your application and start seeing traces in your dashboard

​
Next Steps

Organize with Projects

Step-by-step guide to setting up your projects

Datasets & Data Sources

Connect vector databases to your project

Connect Your AI System

Start sending traces to your project

Collections

Organize traces within a project