Python SDK

This page shows how to call Sakrylle API from Python. Two paths:

• OpenAI compatible: use the official openai package, swap base_url for Sakrylle, call /v1/chat/completions.
• Anthropic native: use the official anthropic package, swap base_url for Sakrylle, call /v1/messages.

Examples are based on the official READMEs pulled via Context7. Refer to the upstream repos for the latest parameters and behavior: openai/openai-python, anthropics/anthropic-sdk-python.

Setup

Put the API key in an environment variable rather than hard-coding:

export SAKRYLLE_API_KEY="sk-xxxxxxxxxxxxxxxx"

OpenAI compatible

Install

pip install openai

Minimal usage

The OpenAI client takes explicit base_url and api_key parameters. Point it at https://api.sakrylle.com/v1 and every call goes through the Sakrylle gateway, never reaching api.openai.com.

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["SAKRYLLE_API_KEY"],
    base_url="https://api.sakrylle.com/v1",
)

completion = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[
        {"role": "system", "content": "You are a concise assistant."},
        {"role": "user", "content": "Explain vector databases in one sentence."},
    ],
)

print(completion.choices[0].message.content)

You can use environment variables instead of explicit arguments:

export OPENAI_API_KEY="$SAKRYLLE_API_KEY"
export OPENAI_BASE_URL="https://api.sakrylle.com/v1"

from openai import OpenAI

client = OpenAI()  # Reads the two env vars automatically

Streaming

stream=True returns an iterator over chunks:

import os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["SAKRYLLE_API_KEY"],
    base_url="https://api.sakrylle.com/v1",
)

stream = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{"role": "user", "content": "Count to 5."}],
    stream=True,
)

for chunk in stream:
    delta = chunk.choices[0].delta.content or ""
    print(delta, end="", flush=True)

Tool calling

Use the tools field exactly as in OpenAI's official API:

import json, os
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["SAKRYLLE_API_KEY"],
    base_url="https://api.sakrylle.com/v1",
)

tools = [
    {
        "type": "function",
        "function": {
            "name": "get_weather",
            "description": "Look up the current weather for a city",
            "parameters": {
                "type": "object",
                "properties": {
                    "city": {"type": "string", "description": "City name"},
                },
                "required": ["city"],
            },
        },
    }
]

resp = client.chat.completions.create(
    model="gpt-5.6-sol",
    messages=[{"role": "user", "content": "What's the weather in Shanghai today?"}],
    tools=tools,
)

choice = resp.choices[0]
if choice.message.tool_calls:
    for call in choice.message.tool_calls:
        print("call:", call.function.name, json.loads(call.function.arguments))
else:
    print(choice.message.content)

Sakrylle does not alter OpenAI's tool calling protocol — whatever the model returns is what you get.

Anthropic native

Install

pip install anthropic

Minimal usage

The Anthropic constructor also accepts base_url:

import os
from anthropic import Anthropic

client = Anthropic(
    api_key=os.environ["SAKRYLLE_API_KEY"],
    base_url="https://api.sakrylle.com",
)

message = client.messages.create(
    model="claude-haiku-4-5-20251001",
    max_tokens=1024,
    messages=[
        {"role": "user", "content": "Introduce Claude in one sentence."},
    ],
)

print(message.content[0].text)

The SDK appends /v1/messages automatically. Use https://api.sakrylle.com as base_url — do not add /v1. The exact joining rules are documented in the anthropic-sdk-python README. The env vars ANTHROPIC_API_KEY and ANTHROPIC_BASE_URL work as alternatives to explicit arguments.

Streaming

messages.stream() is a context manager that yields plain text fragments:

import os
from anthropic import Anthropic

client = Anthropic(
    api_key=os.environ["SAKRYLLE_API_KEY"],
    base_url="https://api.sakrylle.com",
)

with client.messages.stream(
    model="claude-haiku-4-5-20251001",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a four-line poem."}],
) as stream:
    for text in stream.text_stream:
        print(text, end="", flush=True)
    print()

Error handling

Both SDKs use a similar exception hierarchy (AuthenticationError, RateLimitError, APIStatusError, etc.). Error codes align with upstream:

from openai import OpenAI, AuthenticationError, RateLimitError, APIStatusError

client = OpenAI(api_key="bad-key", base_url="https://api.sakrylle.com/v1")

try:
    client.chat.completions.create(
        model="gpt-5.6-sol",
        messages=[{"role": "user", "content": "hi"}],
    )
except AuthenticationError:
    print("Key invalid or revoked")
except RateLimitError:
    print("Rate limited, retry later")
except APIStatusError as e:
    print("Upstream error:", e.status_code, e.response.text)

For a full error table and debugging tips, see Errors.