Core Concepts

This guide explains the fundamental concepts you need to understand when using LangMart. Understanding these concepts will help you make the most of the platform.

Models and Providers

What is a Provider?

A provider is a company that offers AI/LLM services. LangMart integrates with multiple providers, giving you access to their models through a single API.

What is a Model?

A model is a specific AI system trained for language tasks. Each model has unique characteristics:

Model Naming Convention

LangMart uses a standardized naming format:

provider/model-name

Examples:

• openai/gpt-4o - GPT-4o from OpenAI
• anthropic/claude-3.5-sonnet - Claude 3.5 Sonnet from Anthropic
• groq/llama-3.3-70b-versatile - Llama 3.3 70B on Groq infrastructure

Model Categories

Models are organized into categories based on their capabilities:

Connections

What is a Connection?

A connection links LangMart to a provider using your API credentials. Connections enable:

• Using your own provider API keys
• Accessing provider-specific features
• Managing rate limits and quotas
• Tracking usage per connection

Connection Types

Creating a Connection

1. Go to Connections in the sidebar
2. Click Add Connection
3. Select the provider
4. Enter your API key
5. (Optional) Set rate limits and quotas
6. Click Create

Connection Scope

Why Use Your Own Connection?

Credits and Billing

How Billing Works

LangMart uses a credit-based billing system:

1. Purchase Credits - Buy credits through the platform
2. Use Credits - Credits are deducted based on token usage
3. Track Spending - Monitor usage in real-time

Understanding Token Pricing

AI models charge based on tokens - pieces of text roughly equivalent to 4 characters or 3/4 of a word.

Input: "Hello, how are you today?" = ~7 tokens
Output: "I'm doing great, thanks for asking!" = ~9 tokens

Price Calculation

Cost = (Input Tokens * Input Price) + (Output Tokens * Output Price)

Example:

• Model: groq/llama-3.3-70b-versatile
• Input: 1,000 tokens @ $0.0006/1K = $0.0006
• Output: 500 tokens @ $0.0008/1K = $0.0004
• Total: $0.001

Billing Modes

Viewing Usage

Track your usage in multiple ways:

Cost Optimization Tips

1. Choose the Right Model - Don't use GPT-4 when GPT-3.5 suffices
2. Optimize Prompts - Shorter prompts = fewer input tokens
3. Set Max Tokens - Limit output length when appropriate
4. Use Caching - Avoid duplicate requests
5. Monitor Usage - Set up alerts for spending thresholds

Organizations and Teams

What is an Organization?

An organization is a group of users who share resources. Organizations enable:

• Team collaboration
• Shared connections
• Centralized billing
• Access control

Organization Roles

Creating an Organization

1. Go to Organizations in the sidebar
2. Click Create Organization
3. Enter organization name and details
4. Invite team members

Switching Organizations

If you belong to multiple organizations:

1. Click your profile icon
2. Select Switch Organization
3. Choose the organization

Inviting Members

1. Go to Organizations > Your Organization
2. Click Invite Member
3. Enter email address
4. Select role
5. Send invitation

Invitees receive an email with a link to join.

Gateway Architecture (Simplified)

Understanding LangMart's architecture helps you troubleshoot and optimize.

What is a Gateway?

A gateway is a server component that processes your API requests. LangMart uses different gateway types for different purposes.

Request Flow

Your App                LangMart                 Provider
   │                       │                        │
   │──── API Request ─────>│                        │
   │                       │                        │
   │                       │── Validate & Route ───>│
   │                       │                        │
   │                       │<──── Provider API ─────│
   │                       │                        │
   │<─── Response ─────────│                        │

Gateway Types

What This Means for You

• Reliability - Multiple gateways provide redundancy
• Performance - Requests are routed optimally
• Flexibility - Different providers handled appropriately

API Compatibility

OpenAI Compatibility

LangMart's API is designed to be a drop-in replacement for OpenAI's API:

# OpenAI
from openai import OpenAI
client = OpenAI(api_key="sk-...")

# LangMart - just change base_url
from openai import OpenAI
client = OpenAI(
    api_key="sk-langmart-...",
    base_url="https://api.langmart.ai/v1"
)

Supported Endpoints

Provider-Specific Features

Some provider features are available through LangMart:

Alerts and Monitoring

Setting Up Alerts

Alerts notify you when certain conditions are met:

Creating an Alert

1. Go to Alerts in the sidebar
2. Click Create Alert
3. Select alert type and condition
4. Set threshold value
5. Choose notification method
6. Save

Notification Methods

Glossary

Next Steps

Now that you understand the core concepts:

• Create Your First Connection - Add your own provider keys
• Explore Models - Find the right model for your use case
• Set Up Alerts - Monitor your usage
• Manage Your Team - Invite collaborators