Series 1: Getting Started
From zero to your first production AI agent
Getting Started
5 videos
~24 min
From zero to your first production AI agent
This video gives you a high-level overview of why Teleon AI exists and what problems it solves. You'll learn about the three core systems that power every Teleon agent:
Cortex — a persistent memory system with 6 simple methods for storing, searching, and retrieving context across conversations
Helix — an auto-scaling runtime that manages anywhere from 2 to 20+ instances, with built-in health checks and resource management
Sentinel — a production-grade safety layer that handles PII detection, content filtering, and compliance with GDPR and HIPAA
By the end of this video, you'll understand how all three systems work together to produce a fully production-ready agent in just 10 lines of code.
- Introduction
- Three core features
- Hook
2 min
A hands-on quickstart that takes you from zero to a running Teleon agent in under 3 minutes. This video walks through each step in order:
- Installing Teleon from GitHub with a single pip install command
- Configuring your API key securely using environment variables
- Writing your first agent using the @client.agent() decorator
- Running and testing the agent locally
You'll see exactly how Teleon's decorator pattern registers and deploys any function as a production-ready agent — no servers, no config files, no infrastructure setup required.
Prerequisites: Python 3.8+, pip, Git, and a Teleon API key from teleon.ai
- pip install
- API key
- Hello world agent
4 min
View the Teleon library on GitHub
This video addresses the most common limitation of basic AI agents: they have no memory between calls. You'll first see the problem demonstrated live with a simple agent that forgets information from just one message ago, then fix it by enabling Cortex with a single parameter — cortex=True.
- The 4 core Cortex methods: store(), search(), get(), and count()
How semantic search works — matching concepts like "refund" to stored memories tagged "pricing" without any keyword overlap
- A live demo showing memory growing across multiple agent calls
- An overview of the full 6-method Cortex API
- How multi-tenancy keeps each user's memory completely isolated
Key takeaway: One parameter transforms a stateless agent into one that remembers every conversation, understands context across sessions, and maintains separate memory per user.
- Enable Cortex
- Store/search demo
6 min
View the Teleon library on GitHub
A working prototype is not a production agent. This video shows you what it takes to make your agent reliable under real traffic, using Teleon's Helix runtime. Everything is configured directly in the @client.agent() decorator — your agent logic stays completely untouched.
Auto-scaling — set min and max instance counts so your agent scales up under load and back down automatically
- Health checks — automatic detection and recovery from instance failures without manual intervention
- Resource limits — caps on memory and CPU to keep performance stable under any load
- Sentinel guardrails — PII detection and content filtering applied at the runtime level
The video ends with a live walkthrough of the Teleon dev dashboard, where you can monitor all of this in real time.
- Auto-scaling configuration
- Live demo
~6 min
View the Teleon library on GitHub
This video covers Sentinel — Teleon's built-in safety and compliance system — and shows you how to protect your agents from sensitive data exposure without writing any safety code yourself. Configuration happens entirely inside the agent decorator.
- How to enable PII detection for data like emails, phone numbers, and other identifiers
The difference between block mode (reject the request entirely) and redact mode (strip sensitive data and continue)
- How every violation is automatically written to an audit trail for GDPR and HIPAA compliance
A live demo shows both modes in action, so you can decide which behavior fits your use case before you ever ship to production.
- PII detection
- Content filtering demo
6 min
View the Teleon library on GitHub