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.

Privacy Terms

Day 12 · ~12 min●

Yesterday's seen = set() lives in memory. Script ends → set vanishes. Re-run → empty set → all items look new → duplicates.

Real automations survive script ending. The fix: store state somewhere that persists. For Patterns, the state store is a Google Sheet — a row per processed key. Today's pattern:

At start: read all rows from the state Sheet → build seen set.
For each new item: if not in seen, do the work, append a row with the key.
At end: state is durable. Next run reads the same rows back.

python

# Step 1: discover a spreadsheet
found = toolset.execute_action(Action.GOOGLESHEETS_SEARCH_SPREADSHEETS, {"query": ""})
spreadsheet_id = (found.get("files") or found.get("response_data", {}).get("files", []))[0]["id"]

# Step 2: append a marker row
result = toolset.execute_action(Action.GOOGLESHEETS_SHEET_APPEND_GOOGLE_SHEET_ROW, {
    "spreadsheet_id": spreadsheet_id,
    "sheet_name": "Sheet1",
    "values": [["zuzu-day-12-marker"]],
})

# Step 3: confirm the append
updated_range = (result.get("updates") or result.get("response_data", {}).get("updates", {})).get("updatedRange")
print(f"appended to: {updated_range}")

The Sheet is now state. Next time this script runs, it could read rows back and skip the marker.

Where does the spreadsheet come from?

You need at least one spreadsheet in your Drive. The SEARCH_SPREADSHEETS action finds it. Real automations usually create-or-find a specific "state" spreadsheet by name — for today's lesson, taking the first match is fine.

Could I use Notion or a database instead?

Yes — anything you can read and write becomes a state store. Sheets is convenient because it's already in your Drive, easy to inspect manually, and Composio supports both append and read. Tomorrow's lesson uses the same Sheet to count partial failures.

State across runs — the persistent dedup

Why memory isn't enough

Scope	Survives?	When
Local variable	no	Vanishes when function returns
Module-level variable	no	Vanishes when script ends
File on disk	yes (until restart wipes the sandbox)	Survives a single machine run
Sheet / Notion / DB	yes	Survives anywhere — different machine, different day

The sandbox a Patterns lesson runs in is fresh per submission. Nothing on its disk survives. State has to live in a tool.

The two-phase pattern

python

# PHASE 1 — read state
rows = toolset.execute_action(Action.GOOGLESHEETS_BATCH_GET, {
    "spreadsheet_id": ID, "ranges": ["Sheet1!A:A"]
})
seen = {r[0] for r in rows.get("valueRanges", [{}])[0].get("values", []) if r}

# PHASE 2 — process new items, append each as you go
for item in items:
    if item in seen:
        continue
    do_work(item)
    append_row(ID, [item])
    seen.add(item)

Phase 1 builds the seen set from durable storage. Phase 2 mutates the durable storage as it works.

Why append-as-you-go (not write-at-end)

If the script crashes between phase 1's read and a single end-of-loop write, you've done work but haven't recorded it — next run duplicates. Append per item. A crash mid-loop loses only the in-flight item, not all completed work.

Sheet structure for state

The simplest schema: one column.

text

A
├── id-1
├── id-2
├── id-3

More informative: id, timestamp, status.

text

A           B                       C
id          processed_at            status
id-1        2026-05-04T09:00:00Z    ok
id-2        2026-05-04T09:00:01Z    failed

Which you choose depends on what you'll grep for later.

Discovery, not hardcoded IDs

python

found = toolset.execute_action(Action.GOOGLESHEETS_SEARCH_SPREADSHEETS, {"query": ""})
files = found.get("files") or found.get("response_data", {}).get("files", [])
spreadsheet_id = files[0]["id"]

For today's minimum demo: take the first spreadsheet found. Real scripts filter by name (query="name='zuzu-state'") — but the discovery shape is the same.

Defensive response reads

Composio wraps Sheets responses inconsistently:

python

# Append response
updated = (result.get("updates") 
           or result.get("response_data", {}).get("updates", {})).get("updatedRange")

Read both paths. The lesson asserts on whichever exists.

Sign up to practice

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

or

Day 12 · ~12 min●

Yesterday's seen = set() lives in memory. Script ends → set vanishes. Re-run → empty set → all items look new → duplicates.

Real automations survive script ending. The fix: store state somewhere that persists. For Patterns, the state store is a Google Sheet — a row per processed key. Today's pattern:

At start: read all rows from the state Sheet → build seen set.
For each new item: if not in seen, do the work, append a row with the key.
At end: state is durable. Next run reads the same rows back.

python

# Step 1: discover a spreadsheet
found = toolset.execute_action(Action.GOOGLESHEETS_SEARCH_SPREADSHEETS, {"query": ""})
spreadsheet_id = (found.get("files") or found.get("response_data", {}).get("files", []))[0]["id"]

# Step 2: append a marker row
result = toolset.execute_action(Action.GOOGLESHEETS_SHEET_APPEND_GOOGLE_SHEET_ROW, {
    "spreadsheet_id": spreadsheet_id,
    "sheet_name": "Sheet1",
    "values": [["zuzu-day-12-marker"]],
})

# Step 3: confirm the append
updated_range = (result.get("updates") or result.get("response_data", {}).get("updates", {})).get("updatedRange")
print(f"appended to: {updated_range}")

The Sheet is now state. Next time this script runs, it could read rows back and skip the marker.

Where does the spreadsheet come from?

You need at least one spreadsheet in your Drive. The SEARCH_SPREADSHEETS action finds it. Real automations usually create-or-find a specific "state" spreadsheet by name — for today's lesson, taking the first match is fine.

Could I use Notion or a database instead?

Yes — anything you can read and write becomes a state store. Sheets is convenient because it's already in your Drive, easy to inspect manually, and Composio supports both append and read. Tomorrow's lesson uses the same Sheet to count partial failures.

State across runs — the persistent dedup

Why memory isn't enough

Scope	Survives?	When
Local variable	no	Vanishes when function returns
Module-level variable	no	Vanishes when script ends
File on disk	yes (until restart wipes the sandbox)	Survives a single machine run
Sheet / Notion / DB	yes	Survives anywhere — different machine, different day

The sandbox a Patterns lesson runs in is fresh per submission. Nothing on its disk survives. State has to live in a tool.

The two-phase pattern

python

# PHASE 1 — read state
rows = toolset.execute_action(Action.GOOGLESHEETS_BATCH_GET, {
    "spreadsheet_id": ID, "ranges": ["Sheet1!A:A"]
})
seen = {r[0] for r in rows.get("valueRanges", [{}])[0].get("values", []) if r}

# PHASE 2 — process new items, append each as you go
for item in items:
    if item in seen:
        continue
    do_work(item)
    append_row(ID, [item])
    seen.add(item)

Phase 1 builds the seen set from durable storage. Phase 2 mutates the durable storage as it works.

Why append-as-you-go (not write-at-end)

If the script crashes between phase 1's read and a single end-of-loop write, you've done work but haven't recorded it — next run duplicates. Append per item. A crash mid-loop loses only the in-flight item, not all completed work.

Sheet structure for state

The simplest schema: one column.

text

A
├── id-1
├── id-2
├── id-3

More informative: id, timestamp, status.

text

A           B                       C
id          processed_at            status
id-1        2026-05-04T09:00:00Z    ok
id-2        2026-05-04T09:00:01Z    failed

Which you choose depends on what you'll grep for later.

Discovery, not hardcoded IDs

python

found = toolset.execute_action(Action.GOOGLESHEETS_SEARCH_SPREADSHEETS, {"query": ""})
files = found.get("files") or found.get("response_data", {}).get("files", [])
spreadsheet_id = files[0]["id"]

For today's minimum demo: take the first spreadsheet found. Real scripts filter by name (query="name='zuzu-state'") — but the discovery shape is the same.

Defensive response reads

Composio wraps Sheets responses inconsistently:

python

# Append response
updated = (result.get("updates") 
           or result.get("response_data", {}).get("updates", {})).get("updatedRange")

Read both paths. The lesson asserts on whichever exists.

Sign up to practice

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

or