Reca API is a reverse-engineered implementation of Reka AI in OpenAI-compatible API.
- Reverse-engineered Reka AI Playground API
- OpenAI API compatible
- Can be deployed via Docker Compose or run directly
- Supports proxy connections
- Supports three authorization method
- reka-core
- reka-flash
- reka-edge
Clone repo
git clone [email protected]:HuggingBear/Reca-API.git
cd Reca-APIConfigure token or user information in .env file
REKA_TOKEN=YOUR_REKA_JWT_TOKEN (Without 'Bearer ')
REKA_USER=YOUR_USER_NAME
REKA_PASS=YOUR_SUPER_LONG_PASSWORDRun
docker-compose up
- Install dependencies:
pip install -r requirements.txt - Run:
python src/main.py
| Variable | Description |
|---|---|
| PROXY | Proxy address (e.g., "socks5://192.168.0.1:1080") |
| REKA_TOKEN | Access token (JWT token) |
| REKA_USER | Username |
| REKA_PASS | Password |
curl http://127.0.0.1:3031/v1/chat/completions \
-X POST \
-H "Content-Type: application/json" \
-d '{
"model": "chat",
"messages": [
{"role": "user", "content": "Hello, how are you?"}
]
}'Response
{
"id": "chatcmpl-reka-ai",
"object": "chat.completion.chunk",
"created": 1719100000,
"model": "reka-core",
"choices": [
{
"index": 0,
"delta": {
"role": "assistant",
"content": " Hello! 😄 I'm Yasa, a helpful AI assistant. I'm doing great today! How are you doing today?"
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 0,
"completion_tokens": 0,
"total_tokens": 0
}
}You can choose one of the following methods to authenticate:
- Set
REKA_TOKENenvironment variable - Set
REKA_USERandREKA_PASSenvironment variables - Specify
X-Reka-Tokenin request header
This project is a fork of jessfin/reka.
It is not affiliated with Reka AI and may violate their Terms of Service. Use with caution.
This project is licensed under the MIT License.