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.

Few-shot — Ai Foundations | zuzu.codes

Day 11 · ~11 min●

Sentiment classification, but with examples baked into the prompt:

python

prompt = '''Classify each as positive, negative, or neutral.

Examples:
"Best purchase ever" → positive
"Total disappointment" → negative
"It's fine I guess" → neutral

Now classify: "meh, it works"
'''
result = Agent(model).run_sync(prompt)
print(result.output.strip().lower())

The examples teach the model the pattern — three labels, the format →, your classification style. The new input slots into the same shape.

Why is this better than just listing labels?

Two reasons:

Style transfer — "meh" might get classified differently if you'd shown all-caps, exclamation-mark examples. Examples set the tone of what counts as positive/negative.
Format anchor — by demonstrating the → label shape, you nudge the model away from prose like "This is a negative review because...".

Few-shot prompting

Few-shot = include 1-5 examples in the prompt. Zero-shot = no examples, just the directive. The examples teach the model the pattern you want.

The pattern

python

prompt = f'''<task description>

Examples:
<input 1> → <output 1>
<input 2> → <output 2>
<input 3> → <output 3>

Now classify: <new input>
'''

The → is conventional; you can use any consistent separator (:, =>, etc.).

When few-shot helps

Edge-case clarity — examples cover ambiguous cases ("meh" → neutral, not negative)
Format anchoring — model copies the exact output format
Domain-specific style — your team's classification style isn't general-knowledge
Task with no obvious name — "format this as our internal style guide" needs to be shown, not described

When few-shot is overkill

Tasks the model already does well (translation, basic summarization)
One-off prompts where you don't care about format precision
When you can specify the format with words ("reply with one word")

Quality > quantity

3-5 examples is usually enough. More than 5-7 starts diminishing returns. Choose examples that span the range of inputs you'll see (typical case + 1-2 edge cases) rather than 10 typical cases.

Sign up to practice

Create a free account to get started. Paid plans unlock all tracks.

Day 11 · ~11 min●

Sentiment classification, but with examples baked into the prompt:

python

prompt = '''Classify each as positive, negative, or neutral.

Examples:
"Best purchase ever" → positive
"Total disappointment" → negative
"It's fine I guess" → neutral

Now classify: "meh, it works"
'''
result = Agent(model).run_sync(prompt)
print(result.output.strip().lower())

The examples teach the model the pattern — three labels, the format →, your classification style. The new input slots into the same shape.

Why is this better than just listing labels?

Two reasons:

Style transfer — "meh" might get classified differently if you'd shown all-caps, exclamation-mark examples. Examples set the tone of what counts as positive/negative.
Format anchor — by demonstrating the → label shape, you nudge the model away from prose like "This is a negative review because...".

Few-shot prompting

Few-shot = include 1-5 examples in the prompt. Zero-shot = no examples, just the directive. The examples teach the model the pattern you want.

The pattern

python

prompt = f'''<task description>

Examples:
<input 1> → <output 1>
<input 2> → <output 2>
<input 3> → <output 3>

Now classify: <new input>
'''

The → is conventional; you can use any consistent separator (:, =>, etc.).

When few-shot helps

Edge-case clarity — examples cover ambiguous cases ("meh" → neutral, not negative)
Format anchoring — model copies the exact output format
Domain-specific style — your team's classification style isn't general-knowledge
Task with no obvious name — "format this as our internal style guide" needs to be shown, not described

When few-shot is overkill

Tasks the model already does well (translation, basic summarization)
One-off prompts where you don't care about format precision
When you can specify the format with words ("reply with one word")

Quality > quantity

3-5 examples is usually enough. More than 5-7 starts diminishing returns. Choose examples that span the range of inputs you'll see (typical case + 1-2 edge cases) rather than 10 typical cases.