Turn Kiro into an OpenAI-compatible and Anthropic-compatible API with one click
Powered by https://github.com/jwadow/kiro-gateway
๐บ๐ธ English โข ๐จ๐ณ ไธญๆ โข ๐ฏ๐ต ๆฅๆฌ่ช โข ๐ฐ๐ท ํ๊ตญ์ด โข ๐ท๐บ ะ ัััะบะธะน โข ๐ช๐ธ Espaรฑol โข ๐ง๐ท Portuguรชs โข ๐ฎ๐ฉ Indonesia
![]() |
Thanks to ai18n for sponsoring this project! ai18n is a reliable and efficient API relay service provider, offering relay Claude models for OpenClaw, Claude Code, Codex, Gemini, and more. |
๐ข Notice: The next version will fix compatibility with Claude Code v2.1.69+, which sendsโ Fixedtool_referencecontent blocks insidetool_resultmessages when using the ToolSearch deferred tool mechanism.
KiroaaS (Kiro as a Service) is a desktop gateway that exposes Kiro's AI models through a local OpenAI-compatible and Anthropic-compatible API. Use your favorite AI tools, libraries, and applications with Kiro - no code changes required.
| Feature | Description |
|---|---|
| ๐ OpenAI-compatible API | /v1/chat/completions endpoint for OpenAI SDK |
| ๐ Anthropic-compatible API | /v1/messages endpoint for Anthropic SDK |
| ๐ VPN/Proxy Support | HTTP/SOCKS5 proxy for restricted networks |
| ๐ง Extended Thinking | Reasoning support exclusive to our project |
| ๐๏ธ Vision Support | Send images to model |
| ๐ ๏ธ Tool Calling | Supports function calling |
| ๐ฌ Built-in Chat | Test your setup with the integrated chat interface |
| ๐ก Streaming | Full SSE streaming support |
| ๐ Retry Logic | Automatic retries on errors (403, 429, 5xx) |
| ๐ Smart Token Management | Automatic refresh before expiration |
| ๐ Multi-language UI | English, ไธญๆ, ๆฅๆฌ่ช, ํ๊ตญ์ด, ะ ัััะบะธะน, Espaรฑol, Portuguรชs, Indonesia |
| ๐ CC Switch Integration | One-click import to CC Switch for Claude Code |
| ๐ Auto Update | Built-in update checker keeps you on the latest version |
Download the latest release from GitHub Releases:
| Platform | Architecture | Download |
|---|---|---|
| macOS | Apple Silicon (M1/M2/M3) | KiroaaS_aarch64.dmg |
| macOS | Intel | KiroaaS_x64.dmg |
| Windows | x64 | KiroaaS_x64-setup.exe |
Linux support coming soon.
# Clone the repo
git clone https://github.com/hnewcity/KiroaaS.git
cd KiroaaS
# Install dependencies
npm install
cd python-backend && pip install -r requirements.txt && cd ..
# Run in dev mode
npm run tauri:dev
# Or build for production
npm run tauri:build- Launch KiroaaS
- Configure your Kiro credentials (auto-detected from Kiro CLI if available)
- Generate a Proxy API Key (or use one provided by your administrator)
- Start the server
- Use
http://localhost:8000as your OpenAI/Anthropic API endpoint
curl http://localhost:8000/v1/chat/completions \
-H "Content-Type: application/json" \
-H "Authorization: Bearer YOUR_PROXY_API_KEY" \
-d '{
"model": "claude-sonnet-4-5",
"messages": [{"role": "user", "content": "Hello!"}]
}'from openai import OpenAI
client = OpenAI(
base_url="http://localhost:8000/v1",
api_key="YOUR_PROXY_API_KEY"
)
response = client.chat.completions.create(
model="claude-sonnet-4-5",
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)import OpenAI from 'openai';
const client = new OpenAI({
baseURL: 'http://localhost:8000/v1',
apiKey: 'YOUR_PROXY_API_KEY',
});
const response = await client.chat.completions.create({
model: 'claude-sonnet-4-5',
messages: [{ role: 'user', content: 'Hello!' }],
});
console.log(response.choices[0].message.content);KiroaaS is compatible with popular AI tools and libraries:
| Category | Tools |
|---|---|
| Python | OpenAI SDK, Anthropic SDK, LangChain, LlamaIndex |
| JavaScript | OpenAI Node.js SDK, Anthropic SDK, Vercel AI SDK |
| IDE Extensions | Cursor, Continue, Cline, Claude Code |
| Chat Apps | ChatGPT-Next-Web, LobeChat, Open WebUI |
KiroaaS supports multiple authentication methods:
| Method | Description |
|---|---|
| Kiro CLI Database | Auto-detect credentials from Kiro CLI (recommended) |
| Credentials File | Use a JSON credentials file |
| Refresh Token | Manually enter your refresh token |
| Option | Default | Description |
|---|---|---|
| Host | 127.0.0.1 |
Server bind address |
| Port | 8000 |
Server port |
| Proxy API Key | - | Required key for API authentication |
| Option | Description |
|---|---|
| Show Thinking Process | Toggle whether thinking/reasoning output is exposed through the compatible APIs |
| VPN/Proxy URL | HTTP/SOCKS5 proxy for network restrictions |
| First Token Timeout | Timeout for initial response (seconds) |
| Streaming Read Timeout | Timeout for streaming responses (seconds) |
| Layer | Technology |
|---|---|
| Frontend | React + TypeScript + Tailwind CSS |
| Desktop | Tauri (Rust) |
| Backend | Python + FastAPI |
Contributions are welcome! Feel free to:
- ๐ Report bugs via GitHub Issues
- ๐ก Suggest features
- ๐ง Submit pull requests
- ๐ Help with translations
AGPL-3.0 ยฉ KiroaaS Contributors

