Skip to main content

​
What You’ll Learn

  • Sending prompts to the Google Gemini API (gemini-2.0-flash) using the google-genai SDK
  • Using client.models.generate_content() — the newer Google AI SDK interface
  • Stripping markdown code fences from LLM responses before execution
  • Writing generated code into a Declaw sandbox filesystem
  • Executing the code securely with sbx.commands.run()
  • Graceful demo mode when no API key is configured

​
Prerequisites

  • Declaw instance running and DECLAW_API_KEY / DECLAW_DOMAIN set
  • GOOGLE_API_KEY (optional — the example runs in demo mode without it)
pip install declaw python-dotenv google-genai

​
Code Walkthrough

This example is available in Python. TypeScript support coming soon.

​
1. Ask Gemini to generate Python code

The google-genai SDK (the newer Google AI SDK, distinct from the older google-generativeai) uses a Client with models.generate_content(): 
from google import genai
from declaw import Sandbox

client = genai.Client()

response = client.models.generate_content(
    model="gemini-2.0-flash",
    contents=(
        "You are a Python code interpreter. When asked a question, "
        "respond ONLY with Python code that computes the answer and "
        "prints it. No markdown, no explanation, just code.\n\n"
        "Question: Count the frequency of each word in "
        "'the quick brown fox jumps over the lazy dog the fox the dog'"
    ),
)
Use google-genai (not google-generativeai). The newer SDK ships as from google import genai and uses the genai.Client() pattern shown above.

​
2. Strip code fences and execute in a sandbox

def strip_code_fences(code: str) -> str:
    code = code.strip()
    if code.startswith("```"):
        code = "\n".join(code.split("\n")[1:])
    if code.endswith("```"):
        code = "\n".join(code.split("\n")[:-1])
    return code.strip()

code = strip_code_fences(response.text or "")

sbx = Sandbox.create(template="python", timeout=300)
try:
    sbx.files.write("/tmp/solution.py", code)
    result = sbx.commands.run("python3 /tmp/solution.py", timeout=30)
    print(result.stdout)
    print(result.exit_code)
finally:
    sbx.kill()

​
3. Demo mode (no API key needed)

code = """\
from collections import Counter

text = "the quick brown fox jumps over the lazy dog the fox the dog"
words = text.split()
counter = Counter(words)

print("Word frequencies:")
for word, count in counter.most_common():
    print(f"  {word}: {count}")
print(f"\\nTotal words: {len(words)}")
print(f"Unique words: {len(counter)}")
"""

sbx = Sandbox.create(template="python", timeout=300)
try:
    sbx.files.write("/tmp/demo.py", code)
    result = sbx.commands.run("python3 /tmp/demo.py", timeout=30)
    print(result.stdout)
finally:
    sbx.kill()

​
Expected Output

============================================================
Gemini Code Interpreter with Declaw Sandbox
============================================================

--- Demo: Running pre-written code in Declaw sandbox ---

Sandbox created: sbx_abc123
  Output:
Word frequencies:
  the: 4
  fox: 2
  dog: 2
  quick: 1
  brown: 1
  jumps: 1
  over: 1
  lazy: 1

Total words: 12
Unique words: 8

Sandbox killed.