Code Interpreter - ZeroTwo Docs

What is Code Interpreter?

Code Interpreter is a powerful tool that allows AI to write and execute Python code in a secure sandbox environment. It enables complex calculations, data analysis, file manipulation, and more - all within your conversation.

Think of Code Interpreter as giving AI the ability to “show its work” by writing and running actual code to solve problems.

When to Use Code Interpreter

Data Analysis

Analyze CSV files, calculate statistics, identify trends

Mathematical Operations

Complex calculations, equations, mathematical modeling

File Processing

Convert formats, extract data, manipulate files

Visualization

Create charts, graphs, and data visualizations

Problem Solving

Algorithm implementation, logic verification

Simulations

Run simulations, test scenarios, model systems

Enabling Code Interpreter

Per Message
Always On
In Projects

Enable for specific tasks:

Click Code Interpreter toggle in top bar
Toggle turns blue when active
Send your request
AI will use code when beneficial

AI automatically decides when to use code based on your request.

How It Works

Request Analysis

AI determines if code execution would help answer your question

Code Generation

AI writes Python code to solve the problem

import pandas as pd
import matplotlib.pyplot as plt

# AI writes relevant code
data = pd.read_csv('sales_data.csv')
total_sales = data['revenue'].sum()

Code Execution

Code runs in secure sandbox environment:

Isolated from your system
No internet access
Time-limited execution
Safe and contained

Results Returned

AI receives output and incorporates it into response:

Calculation results
Generated visualizations
Error messages (if any)
Data outputs

Explanation

AI explains the code and results in natural language

What You Can Do

Data Analysis

CSV Analysis
Statistical Analysis
Data Cleaning
Trend Analysis

Analyze uploaded CSV files:

Analyze this sales data and tell me:
1. Total revenue
2. Best performing month
3. Average order value
4. Revenue trend over time

[upload sales_data.csv]

AI will:

Read and parse the CSV
Calculate requested metrics
Identify patterns
Create visualizations

Calculate statistics:

Calculate mean, median, mode, and standard deviation 
for the price column in this dataset

# AI-generated code
stats = {
    'mean': data['price'].mean(),
    'median': data['price'].median(),
    'mode': data['price'].mode()[0],
    'std': data['price'].std()
}

Clean and prepare data:

Clean this dataset:
- Remove duplicates
- Handle missing values
- Standardize date format
- Remove outliers

AI writes code to clean and export cleaned data.

Identify patterns:

Plot the trend of website visits over the past 6 months
and identify any seasonal patterns

Creates visualizations and analysis.

Mathematical Operations

Complex Calculations
Matrix Operations
Financial Calculations

Solve equations:

Calculate compound interest on $10,000 invested at 
5% annual rate, compounded quarterly for 10 years

principal = 10000
rate = 0.05
times_compounded = 4
years = 10

amount = principal * (1 + rate/times_compounded)**(times_compounded*years)
interest = amount - principal

Linear algebra:

Multiply these two matrices and find the determinant
Matrix A: [[1,2],[3,4]]
Matrix B: [[5,6],[7,8]]

Uses NumPy for matrix operations.

Finance and accounting:

Calculate NPV for a project with initial investment 
of $100k and cash flows of $30k, $40k, $50k over 
3 years at 10% discount rate

Shows full calculation with explanation.

Visualizations

Charts & Graphs
Data Plots
Heatmaps

Create visualizations:

Create a bar chart comparing revenue across regions

import matplotlib.pyplot as plt

regions = ['North', 'South', 'East', 'West']
revenue = [45000, 38000, 52000, 41000]

plt.bar(regions, revenue)
plt.title('Revenue by Region')
plt.xlabel('Region')
plt.ylabel('Revenue ($)')
plt.show()

Chart appears directly in chat.

Line plots, scatter plots:

Plot the relationship between advertising spend 
and sales revenue

Creates scatter plot with trend line.

Correlation analysis:

Create a heatmap showing correlations between 
all numeric columns in the dataset

Visual correlation matrix.

File Operations

Format Conversion
Data Extraction
File Generation

Convert between formats:

Convert this CSV to JSON format

Convert this JSON to a formatted Excel file

Downloads converted file.

Extract specific data:

Extract all email addresses from this text file

Pull out dates and amounts from these receipts

Returns structured data.

Create files:

Generate a CSV file with 100 sample customer 
records including name, email, and purchase history

Creates and provides download link.

Understanding Code Output

Reading Results

Code Block Display

AI shows the code it’s executing:

# Calculating total sales
df = pd.read_csv('sales.csv')
total = df['amount'].sum()
print(f"Total sales: ${total:,.2f}")

Syntax highlighted
Comments explaining steps
Copyable with one click

Execution Output

Results appear in output block:

Total sales: $1,234,567.89

Top 5 products:
Widget A - $450,233
Gadget B - $389,012
Tool C - $278,901
Device D - $234,567
Gizmo E - $198,432

Clear, formatted output.

Generated Images

Charts and graphs display inline:

![Example chart showing data visualization]

High-quality images
Downloadable
Regenerate with different styles

Error Messages

If code fails:

Error: KeyError: 'revenue'
Column 'revenue' not found in dataframe.
Available columns: ['date', 'product', 'sales_amount']

Clear error description
What went wrong
AI suggests fix

Available Python Libraries

Code Interpreter includes many popular libraries:

Data Analysis

import pandas as pd           # Data manipulation
import numpy as np            # Numerical computing
from scipy import stats       # Statistical functions

Visualization

import matplotlib.pyplot as plt  # Plotting
import seaborn as sns            # Statistical viz

Machine Learning

from sklearn import *         # Scikit-learn

Utilities

import json                   # JSON handling
import csv                    # CSV operations
import datetime               # Date/time
import re                     # Regular expressions
import math                   # Mathematical functions

Most popular Python libraries are available. If you need something specific, just ask!

Advanced Features

Multi-Step Operations

Complex workflows with multiple code executions:

Load and Clean

# Load data
df = pd.read_csv('data.csv')
# Clean data
df = df.dropna()
df['date'] = pd.to_datetime(df['date'])

Transform

# Calculate metrics
df['month'] = df['date'].dt.month
monthly_sales = df.groupby('month')['revenue'].sum()

Visualize

# Create chart
plt.plot(monthly_sales.index, monthly_sales.values)
plt.title('Monthly Revenue Trend')
plt.show()

Export

# Save results
monthly_sales.to_csv('monthly_summary.csv')

Interactive Analysis

Have conversations about the code:

You: Analyze this dataset for trends

AI: [Runs analysis code]
Here's what I found...

You: Can you break down the results by region?

AI: [Modifies code to add regional breakdown]
Sure, here's the regional analysis...

You: Now create a visualization

AI: [Generates chart code]
Here's the visualization...

Error Recovery

AI handles errors gracefully:

Code fails → AI sees error
AI analyzes → Understands problem
AI fixes → Rewrites code
Code runs → Success!

You see the fix and explanation.

Security & Limitations

Sandbox Environment

Isolated Execution

Code runs completely isolated from your system

No Network Access

Cannot make external network requests

Time Limits

Execution timeout prevents infinite loops

Memory Limits

Limited RAM prevents resource exhaustion

What’s Allowed

✅ File operations (uploaded files only)
✅ Data analysis and calculations
✅ Image generation (charts/graphs)
✅ Mathematical operations
✅ String manipulation
✅ Data transformations

What’s Not Allowed

❌ Network requests
❌ System file access
❌ Installing packages
❌ Long-running operations (>60s)
❌ Large file generation (>50MB)
❌ Cryptocurrency mining

Code Interpreter is powerful but sandboxed for security. It cannot access your local files or make network requests.

Best Practices

Be Specific About Goals

Better requests:❌ “Analyze this data”
✅ “Calculate the average, median, and identify the top 10 highest values in the ‘sales’ column”Specificity helps AI write targeted code.

Provide Context

Include relevant information:

This CSV contains sales data with columns:
- date (YYYY-MM-DD format)
- product_id
- quantity
- price
- region

Calculate total revenue by region

Context prevents errors.

Start Simple

Build complexity gradually:

First: “Load and show first 5 rows”
Then: “Calculate summary statistics”
Finally: “Create detailed visualization”

Step-by-step is more reliable.

Verify Results

Check important calculations:

Spot check numbers
Verify logic makes sense
Compare with expected ranges
Ask AI to explain unusual results

AI is powerful but can make mistakes.

Example Use Cases

Sales Analysis

Analyze this sales data:
Total revenue by month
Best performing products
Growth rate month-over-month
Create a line chart of revenue trend
Identify any anomalies

[upload sales_data.csv]

Scientific Calculations

Calculate the trajectory of a projectile launched at:
- Initial velocity: 50 m/s
- Angle: 45 degrees
- Height: 10 meters

Show:
- Maximum height reached
- Total distance traveled
- Time of flight
- Plot the trajectory

Data Transformation

Transform this customer data:
Convert dates to ISO format
Standardize phone numbers
Clean up email addresses
Remove duplicates
Export as clean CSV

Troubleshooting

Code Not Executing

Possible causes:

Code Interpreter not enabled
Syntax error in code
Execution timeout

Solutions:

Verify toggle is on
Ask AI to fix errors
Simplify the operation

Wrong Results

Common issues:

Misunderstood requirements
Wrong assumptions about data
Logic error in code

Solutions:

Clarify your requirements
Provide data structure details
Ask AI to explain its approach
Verify step by step

File Not Found

Check:

File was uploaded successfully
Correct filename used
File format supported

Solutions:

Re-upload file
Check filename spelling
Convert to supported format

Pricing & Limits

Free Tier

10 code executions per day
30 second timeout
Standard Python libraries
10MB max file size

Pro Tier

100 code executions per day
60 second timeout
All Python libraries
50MB max file size

Team Tier

Unlimited executions
120 second timeout
Custom library requests
100MB max file size

Execution limits reset daily at midnight UTC.

Next Steps

Deep Research

Combine with research for data-driven insights

Canvas

Edit code in Code Canvas for larger projects

File Uploads

Learn about file upload capabilities

Web Search

Combine with real-time data retrieval

Code Interpreter transforms AI from answering questions to solving problems with actual computation!

​What is Code Interpreter?

​When to Use Code Interpreter

Data Analysis

Mathematical Operations

File Processing

Visualization

Problem Solving

Simulations

​Enabling Code Interpreter

​How It Works

​What You Can Do

​Data Analysis

​Mathematical Operations

​Visualizations

​File Operations

​Understanding Code Output

​Reading Results

​Available Python Libraries

​Data Analysis

​Visualization

​Machine Learning

​Utilities

​Advanced Features

​Multi-Step Operations

​Interactive Analysis

​Error Recovery

​Security & Limitations

​Sandbox Environment

Isolated Execution

No Network Access

Time Limits

Memory Limits

​What’s Allowed

​What’s Not Allowed

​Best Practices

​Example Use Cases

​Sales Analysis

​Scientific Calculations

​Data Transformation

​Troubleshooting

​Pricing & Limits

​Free Tier

​Pro Tier

​Team Tier

​Next Steps

Deep Research

Canvas

File Uploads

Web Search

What is Code Interpreter?

When to Use Code Interpreter

Enabling Code Interpreter

How It Works

What You Can Do

Data Analysis

Mathematical Operations

Visualizations

File Operations

Understanding Code Output

Reading Results

Available Python Libraries

Data Analysis

Visualization

Machine Learning

Utilities

Advanced Features

Multi-Step Operations

Interactive Analysis

Error Recovery

Security & Limitations

Sandbox Environment

What’s Allowed

What’s Not Allowed

Best Practices

Example Use Cases

Sales Analysis

Scientific Calculations

Data Transformation

Troubleshooting

Pricing & Limits

Free Tier

Pro Tier

Team Tier

Next Steps