How to beginner · 3 min read

How to use stop sequences in Claude API

Q: How to use stop sequences in Claude API

Use the stop_sequences parameter in the client.messages.create call with the Anthropic SDK to specify one or more strings where Claude should stop generating text. This parameter accepts a list of strings, and generation halts when any of these sequences is encountered.

Quick answer

Use the stop_sequences parameter in the client.messages.create call with the Anthropic SDK to specify one or more strings where Claude should stop generating text. This parameter accepts a list of strings, and generation halts when any of these sequences is encountered.

PREREQUISITES

Python 3.8+
Anthropic API key
pip install anthropic>=0.20

Setup

Install the Anthropic Python SDK and set your API key as an environment variable before running the code.

Install SDK: pip install anthropic>=0.20
Set environment variable: export ANTHROPIC_API_KEY='your_api_key_here' (Linux/macOS) or set ANTHROPIC_API_KEY=your_api_key_here (Windows)

bash

pip install anthropic>=0.20

Step by step

This example demonstrates how to use the stop_sequences parameter to stop Claude's output when a specific token or phrase is generated.

python

import os
import anthropic

client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])

response = client.messages.create(
    model="claude-3-5-haiku-20241022",
    max_tokens=100,
    system="You are a helpful assistant.",
    messages=[{"role": "user", "content": "Write a short poem about spring."}],
    stop_sequences=["\nHuman:"]
)

print(response.content[0].text)

output

Write a short poem about spring.

Spring wakes the earth with gentle breeze,
Soft petals dance on blooming trees,
Birds sing songs of bright new days,
Life returns in vibrant ways.

Common variations

You can specify multiple stop sequences as a list to halt generation on any of them. The stop_sequences parameter works with all Claude models. For asynchronous usage, use asyncio with the Anthropic client. Adjust max_tokens and system prompts as needed.

python

import os
import asyncio
import anthropic

async def main():
    client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])
    response = await client.messages.acreate(
        model="claude-3-5-haiku-20241022",
        max_tokens=100,
        system="You are a helpful assistant.",
        messages=[{"role": "user", "content": "Explain photosynthesis briefly."}],
        stop_sequences=["\nHuman:", "\nAssistant:"]
    )
    print(response.content[0].text)

asyncio.run(main())

output

Photosynthesis is the process by which green plants use sunlight to convert carbon dioxide and water into glucose and oxygen, providing energy for growth.

Troubleshooting

If the model does not stop where expected, verify your stop_sequences exactly match the output tokens including whitespace and punctuation. Increase max_tokens if the output is cut off prematurely. Check for API errors or rate limits in your environment.

✅

Key Takeaways

Use the stop_sequences parameter to control where Claude stops generating text.
Pass a list of strings to stop_sequences to specify multiple stopping points.
Ensure stop sequences exactly match expected output tokens including whitespace.
The Anthropic SDK supports both synchronous and asynchronous calls with stop sequences.
Adjust max_tokens to avoid premature truncation of responses.

Verified 2026-04 · claude-3-5-haiku-20241022

Verify ↗