Getting Started

Set up Kalibr in 5 minutes.

You will learn:

• How to install the SDK
• How to track your first LLM call
• How to add workflow and customer context

Step 1: Install

pip install kalibr

Step 2: Get Your API Key

1. Go to dashboard.kalibr.systems
2. Sign up or log in
3. Create an organization when prompted
4. Click your email (top right) → API Keys → Create
5. Copy your API key (starts with sk_)

Step 3: Set Environment Variables

export KALIBR_API_KEY=sk_...
export KALIBR_COLLECTOR_URL=https://api.kalibr.systems/api/ingest

Both are required. KALIBR_API_KEY authenticates you. KALIBR_COLLECTOR_URL tells the SDK where to send traces.

Step 4: Add One Line

import kalibr  # Add this line FIRST
import openai

client = openai.OpenAI()
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello"}]
)

Go to dashboard.kalibr.systems. You'll see your trace.

Adding Context

Track Workflows

Group related calls into a workflow:

import os
os.environ["KALIBR_WORKFLOW_ID"] = "research_pipeline"

Track Customers

Attribute costs to specific customers:

import os
os.environ["KALIBR_TENANT_ID"] = "acme_corp"

Track Non-LLM Operations

Use the @trace decorator:

from kalibr import trace

@trace(operation="fetch_docs", provider="database", model="postgres")
def fetch_documents(query):
    return db.query(query)

Verify Setup

curl -H "X-API-Key: $KALIBR_API_KEY" https://api.kalibr.systems/health

Expected: {"status": "ok"}

Troubleshooting

Next Steps

• Concepts — Understand the data model
• Workflows — Track multi-step operations
• Traces — Debug individual executions