How much does zuzu.codes cost?

The starter track is free — read all lessons and practice for free. Full access to every track (current and future) is $14.99/month. Cancel anytime.

How long does each track take?

Each track is designed as a 30-day challenge — one lesson per day, about 15 minutes each. Go at your own pace, but the structure is built around daily consistency.

What's the lesson format?

Each lesson is a student-teacher dialogue with code examples, followed by a hands-on code challenge in an in-browser editor. You read, you understand, then you write real code.

Do I need prior coding experience?

Our beginner track starts from absolute zero — no prior experience needed. Advanced tracks build on earlier ones, and the platform tells you exactly where to start.

How is zuzu.codes different from freeCodeCamp or Codecademy?

zuzu.codes uses a structured 30-day track format with dialogue-based teaching, an in-browser code editor, and gamification (XP, streaks, progress tracking). The format builds genuine understanding through daily practice.

Multi-turn context — Ai Foundations

Day 17 · ~11 min●

Yesterday you ran 2 turns. Today: 3, with the model accumulating context across all of them.

python

from pydantic_ai import Agent
agent = Agent(model)

turns = [
    "What's the capital of France?",
    "What about Germany?",
    "Are these in the same continent?",
]

history = []
for q in turns:
    result = agent.run_sync(q, message_history=history)
    print(f"USER: {q}")
    print(f"ASSISTANT: {result.output}\n")
    history = result.all_messages()

Turn 3's question — "are these in the same continent?" — is meaningless without turns 1 and 2. The model resolves "these" by looking at the prior turns it was given.

Multi-turn — letting the model see the conversation

The pattern from yesterday, generalized to N turns:

python

history = []
for user_input in turns:
    result = agent.run_sync(user_input, message_history=history)
    history = result.all_messages()    # extend history for next turn
    use(result.output)

Each iteration:

Send this turn's user input + accumulated history
Get the model's response
Update history to include both the new user input and the model's response
Use the response (print, save, route, etc.)

Reference resolution

The payoff of multi-turn is contextual references:

"What about Germany?" — Germany is a fresh noun, but What about only makes sense relative to a prior question
"Are these in the same continent?" — these resolves to the entities from turns 1-2
"Tell me more about that" — that refers to the most recent topic

Without history, the model would be confused. With history, it follows naturally.

When history is hurting

Long histories cost tokens (you re-send everything every turn) and can confuse the model when irrelevant context bleeds into a new topic. Two recovery strategies:

Reset for new topic — when the user pivots, drop the old history
Summarize old turns — replace the first 5 turns with a single summary message before continuing

For week-3 lessons we keep histories short and don't worry about trimming. Production multi-turn (chatbots, agents) handles this — AI Intermediate.

Roles

The message list has two roles:

Role	Source
`user`	Each `run_sync(prompt)` adds the prompt as a user message
`assistant`	Each model reply is added as an assistant message

A third role, system, sets the model's persona and rules. That's day 17.

Day 17 · ~11 min●

Yesterday you ran 2 turns. Today: 3, with the model accumulating context across all of them.

python

from pydantic_ai import Agent
agent = Agent(model)

turns = [
    "What's the capital of France?",
    "What about Germany?",
    "Are these in the same continent?",
]

history = []
for q in turns:
    result = agent.run_sync(q, message_history=history)
    print(f"USER: {q}")
    print(f"ASSISTANT: {result.output}\n")
    history = result.all_messages()

Turn 3's question — "are these in the same continent?" — is meaningless without turns 1 and 2. The model resolves "these" by looking at the prior turns it was given.

Multi-turn — letting the model see the conversation

The pattern from yesterday, generalized to N turns:

python

history = []
for user_input in turns:
    result = agent.run_sync(user_input, message_history=history)
    history = result.all_messages()    # extend history for next turn
    use(result.output)

Each iteration:

Send this turn's user input + accumulated history
Get the model's response
Update history to include both the new user input and the model's response
Use the response (print, save, route, etc.)

Reference resolution

The payoff of multi-turn is contextual references:

"What about Germany?" — Germany is a fresh noun, but What about only makes sense relative to a prior question
"Are these in the same continent?" — these resolves to the entities from turns 1-2
"Tell me more about that" — that refers to the most recent topic

Without history, the model would be confused. With history, it follows naturally.

When history is hurting

Long histories cost tokens (you re-send everything every turn) and can confuse the model when irrelevant context bleeds into a new topic. Two recovery strategies:

Reset for new topic — when the user pivots, drop the old history
Summarize old turns — replace the first 5 turns with a single summary message before continuing

For week-3 lessons we keep histories short and don't worry about trimming. Production multi-turn (chatbots, agents) handles this — AI Intermediate.

Roles

The message list has two roles:

Role	Source
`user`	Each `run_sync(prompt)` adds the prompt as a user message
`assistant`	Each model reply is added as an assistant message

A third role, system, sets the model's persona and rules. That's day 17.

Multi-turn — letting the model see the conversation

Reference resolution

When history is hurting

Roles

Multi-turn — letting the model see the conversation

Reference resolution

When history is hurting

Roles

Multi-turn — letting the model see the conversation

Reference resolution

When history is hurting

Roles

Sign up to practice

Multi-turn — letting the model see the conversation

Reference resolution

When history is hurting

Roles

Sign up to practice