• Python 3.10+
• Kiro IDE with logged in account

# Clone the repository
git clone https://github.com/Jwadow/kiro-openai-gateway.git
cd kiro-openai-gateway

# Install dependencies
pip install -r requirements.txt

# Configure (see Configuration section)
cp .env.example .env
# Edit .env with your credentials

# Start the server
python main.py

The server will be available at http://localhost:8000

Specify the path to the credentials file:

KIRO_CREDS_FILE="~/.aws/sso/cache/kiro-auth-token.json"

# Password to protect YOUR proxy server (make up any secure string)
# You'll use this as api_key when connecting to your gateway
PROXY_API_KEY="my-super-secret-password-123"

{
  "accessToken": "eyJ...",
  "refreshToken": "eyJ...",
  "expiresAt": "2025-01-12T23:00:00.000Z",
  "profileArn": "arn:aws:codewhisperer:us-east-1:...",
  "region": "us-east-1"
}

Create a .env file in the project root:

# Required
REFRESH_TOKEN="your_kiro_refresh_token"

# Password to protect YOUR proxy server (make up any secure string)
PROXY_API_KEY="my-super-secret-password-123"

# Optional
PROFILE_ARN="arn:aws:codewhisperer:us-east-1:..."
KIRO_REGION="us-east-1"

The refresh token can be obtained by intercepting Kiro IDE traffic. Look for requests to:

• prod.us-east-1.auth.desktop.kiro.dev/refreshToken

curl http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer my-super-secret-password-123" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "Hello!"}],
    "stream": true
  }'

curl http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer my-super-secret-password-123" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "What is 2+2?"}
    ],
    "stream": true
  }'

curl http://localhost:8000/v1/chat/completions \
  -H "Authorization: Bearer my-super-secret-password-123" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-5",
    "messages": [{"role": "user", "content": "What is the weather in London?"}],
    "tools": [{
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get weather for a location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {"type": "string", "description": "City name"}
          },
          "required": ["location"]
        }
      }
    }]
  }'

from openai import OpenAI

client = OpenAI(
    base_url="http://localhost:8000/v1",
    api_key="my-super-secret-password-123"  # Your PROXY_API_KEY from .env
)

response = client.chat.completions.create(
    model="claude-sonnet-4-5",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"}
    ],
    stream=True
)

for chunk in response:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="")

from langchain_openai import ChatOpenAI

llm = ChatOpenAI(
    base_url="http://localhost:8000/v1",
    api_key="my-super-secret-password-123",  # Your PROXY_API_KEY from .env
    model="claude-sonnet-4-5"
)

response = llm.invoke("Hello, how are you?")
print(response.content)

kiro-openai-gateway/
├── main.py                    # Entry point, FastAPI app creation
├── requirements.txt           # Python dependencies
├── .env.example               # Environment configuration example
│
├── kiro_gateway/              # Main package
│   ├── __init__.py            # Package exports
│   ├── config.py              # Configuration and constants
│   ├── models.py              # Pydantic models for OpenAI API
│   ├── auth.py                # KiroAuthManager - token management
│   ├── cache.py               # ModelInfoCache - model caching
│   ├── utils.py               # Helper utilities
│   ├── converters.py          # OpenAI <-> Kiro conversion
│   ├── parsers.py             # AWS SSE stream parsers
│   ├── streaming.py           # Response streaming logic
│   ├── http_client.py         # HTTP client with retry logic
│   ├── debug_logger.py        # Debug logging (optional)
│   └── routes.py              # FastAPI routes
│
├── tests/                     # Tests
│   ├── unit/                  # Unit tests
│   └── integration/           # Integration tests
│
└── debug_logs/                # Debug logs (generated when enabled)

Debug logging is disabled by default. To enable, add to your .env:

DEBUG_LAST_REQUEST=true

When enabled, requests are logged to the debug_logs/ folder:

# Run all tests
pytest

# Run unit tests only
pytest tests/unit/

# Run with coverage
pytest --cov=kiro_gateway

The modular architecture makes it easy to add support for other providers:

1. Create a new module kiro_gateway/providers/new_provider.py
2. Implement the required classes:
   • NewProviderAuthManager — token management
   • NewProviderConverter — format conversion
   • NewProviderParser — response parsing
3. Add routes to routes.py or create a separate router

This project is licensed under the GNU Affero General Public License v3.0 (AGPL-3.0).

This means:

• ✅ You can use, modify, and distribute this software
• ✅ You can use it for commercial purposes
• ⚠️ You must disclose source code when you distribute the software
• ⚠️ Network use is distribution — if you run a modified version on a server and let others interact with it, you must make the source code available to them
• ⚠️ Modifications must be released under the same license

See the LICENSE file for the full license text.

AGPL-3.0 ensures that improvements to this software benefit the entire community. If you modify this gateway and deploy it as a service, you must share your improvements with your users.

By submitting a contribution to this project, you agree to the terms of our Contributor License Agreement (CLA). This ensures that:

• You have the right to submit the contribution
• You grant the maintainer rights to use and relicense your contribution
• The project remains legally protected

Jwadow — @Jwadow

This project is not affiliated with, endorsed by, or sponsored by Amazon Web Services (AWS), Anthropic, or Kiro IDE. Use at your own risk and in compliance with the terms of service of the underlying APIs.