Together Code Interpreter (TCI) enables you to execute Python code in a sandboxed environment. At this time, the Code Interpreter only supports Python. In the future, we plan on expanding the language options.

​
Run your first query using the TCI

from together import Together

client = Together()

# Run a simple print statement in the code interpreter
response = client.code_interpreter.run(code='print("Welcome to Together Code Interpreter!")', language="python")

print(f"Status: {response.data.status}")

for output in response.data.outputs:
    print(f"{output.type}: {output.data}")
Output 
Status: completed
stdout: Welcome to Together Code Interpreter!

​
Pricing information

TCI usage is billed at $0.03/session. As detailed below, sessions have a lifespan of 60 minutes and can be used multiple times.TCI is free to use until May 8, 2025.

​
Example Use Cases

Overview of how Together Code Interpreter can be used in a code generation, execution and improvement iteration.
  • Reinforcement learning (RL) training: TCI transforms code execution into an interactive RL environment where generated code is run and evaluated in real time, providing reward signals from successes or failures, integrating automated pass/fail tests, and scaling easily across parallel workers—thus creating a powerful feedback loop that refines coding models over many trials.
  • Developing agentic workflows: TCI allows AI agents to seamlessly write and execute Python code, enabling robust, iterative, and secure computations within a closed-loop system.

​
Response Format

The API returns:
  • session_id: Identifier for the current session
  • outputs: Array of execution outputs, which can include:
    • Execution output (the return value of your snippet)
    • Standard output (stdout)
    • Standard error (stderr)
    • Error messages
    • Rich display data (images, HTML, etc.) Example
{
  "data": {
    "outputs": [
      {
        "data": "Hello, world!\n",
        "type": "stdout"
      },
      {
        "data": {
          "image/png": "iVBORw0KGgoAAAANSUhEUgAAA...",
          "text/plain": "<Figure size 640x480 with 1 Axes>"
        },
        "type": "display_data"
      }
    ],
    "session_id": "ses_CM42NfvvzCab123"
  },
  "errors": null
}

​
Usage overview

Together AI has created sessions to measure TCI usage. A session is an active code execution environment that can be called to execute code, they can be used multiple times and have a lifespan of 60 minutes. Typical TCI usage follows this workflow:
  1. Start a session (create a TCI instance).
  2. Call that session to execute code; TCI outputs stdout and stderr.
  3. Optionally reuse an existing session by calling its session_id.

​
Reusing sessions and maintaining state between runs

The session_id can be used to access a previously initialized session. All packages, variables, and memory will be retained. 
from together import Together

client = Together()

# set a variable x to 42
response1 = client.code_interpreter.run(code="x = 42", language="python")

session_id = response1.data.session_id

# print the value of x
response2 = client.code_interpreter.run(
    code='print(f"The value of x is {x}")', language="python", session_id=session_id
)

for output in response2.data.outputs:
    print(f"{output.type}: {output.data}")
Output 
stdout: The value of x is 42

​
Using the TCI for Data analysis

Together Code Interpreter is a very powerful tool and gives you access to a fully functional coding environment. You can install Python libraries and conduct fully fledged data analysis experiments. 
from together import Together

client = Together()

# Create a code interpreter instance
code_interpreter = client.code_interpreter

code = '''
!pip install numpy
import numpy as np

# Create a random matrix
matrix = np.random.rand(3, 3)
print("Random matrix:")
print(matrix)

# Calculate eigenvalues
eigenvalues = np.linalg.eigvals(matrix)
print("\\nEigenvalues:")
print(eigenvalues)
'''

response = code_interpreter.run(
  	code=code,
    language="python"
)

for output in response.data.outputs:
    print(f"{output.type}: {output.data}")
if response.data.errors:
    print(f"Errors: {response.data.errors}")

​
Uploading and using files with TCI

from together import Together

client = Together()

# Create a code interpreter instance
code_interpreter = client.code_interpreter

script_content = "import sys\nprint(f'Hello from inside {sys.argv[0]}!')"

# Define the script file as a dictionary
script_file = {
    "name": "myscript.py",
    "encoding": "string",
    "content": script_content,
}

code_to_run_script = "!python myscript.py"

response = code_interpreter.run(
    code=code_to_run_script,
    language="python",
    files=[script_file],  # Pass the script dictionary in a list
)

# Print results
print(f"Status: {response.data.status}")
for output in response.data.outputs:
    print(f"{output.type}: {output.data}")
if response.data.errors:
    print(f"Errors: {response.data.errors}")
Output 
Status: completed
stdout: Hello from inside myscript.py!

​
Pre-installed dependencies

TCI’s Python sessions come pre-installed with the following dependencies, any other dependencies can be installed using a !pip install command in the python code. 
- aiohttp
- beautifulsoup4
- bokeh
- gensim
- imageio
- joblib
- librosa
- matplotlib
- nltk
- numpy
- opencv-python
- openpyxl
- pandas
- plotly
- pytest
- python-docx
- pytz
- requests
- scikit-image
- scikit-learn
- scipy
- seaborn
- soundfile
- spacy
- textblob
- tornado
- urllib3
- xarray
- xlrd
- sympy

​
List Active Sessions

To retrieve all your active sessions: 
curl -X GET "https://api.together.ai/tci/sessions" \
  -H "Authorization: Bearer $TOGETHER_API_KEY" \
  -H "Content-Type: application/json"
Output: 
{
  "data": {
  "sessions": [
    {
      "id":"ses_CM3zokEcfkdh5G8UiKApw",
      "started_at":"2025-03-12T10:34:22.125248Z",
      "execute_count":2,
      "last_execute_at":"2025-03-12T10:37:24.145685Z",
      "expires_at":"2025-03-12T11:04:22.125248Z"
    }
    ]
  },
  "errors": null
}

​
Further reading

TCI API Reference docs Together Code Interpreter Notebook

​
Troubleshooting & questions

If you have questions about integrating TCI into your workflow or encounter any issues, please contact us.