Add readme for gai-backend

gai-backend/readme.md

@@ -0,0 +1,119 @@
+# Orchid GenAI Server
+
+This repository contains the components for running an AI model inference server with Orchid nanopayments integration. The system consists of three main components:
+
+- A billing server that handles WebSocket connections and payment processing
+- An inference API that provides a REST interface for model inference
+- A CLI client for testing and interacting with the service
+
+## CLI Client
+
+The CLI client provides a flexible way to interact with the GenAI service, supporting both payment handling and inference requests. It can be run in three modes:
+
+- Normal mode: Handles both payments and inference
+- Wallet-only mode: Just handles payments
+- Inference-only mode: Just makes inference requests using an existing auth token
+
+### Configuration
+
+The client uses a JSON configuration file with the following structure:
+
+```json
+{
+  "inference": {
+    "provider": "default",
+    "funder": "0x...",
+    "secret": "your-wallet-secret",
+    "chainid": 1,
+    "rpc": "https://your-eth-rpc-endpoint"
+  },
+  "location": {
+    "providers": {
+      "default": {
+        "billing_url": "wss://billing-endpoint"
+      }
+    }
+  },
+  "test": {
+    "messages": [
+      {
+        "role": "user",
+        "content": "Hello, how are you?"
+      }
+    ],
+    "model": "model-id",
+    "params": {
+      "temperature": 0.7
+    },
+    "retry_delay": 1.5
+  },
+  "logging": {
+    "level": "INFO",
+    "file": "client.log"
+  }
+}
+```
+
+#### Configuration Sections
+
+- `inference`: Wallet and chain configuration
+  - `provider`: Provider ID matching an entry in the location config
+  - `funder`: Ethereum address of the funding wallet
+  - `secret`: Private key for the funding wallet
+  - `chainid`: Ethereum chain ID
+  - `rpc`: Ethereum RPC endpoint URL
+  - `currency`: (Optional) Currency for payments
+
+- `location`: Service endpoint configuration
+  - `providers`: Map of provider IDs to their connection details
+    - `billing_url`: WebSocket URL for the billing server
+
+- `test`: Test configuration for inference requests
+  - `messages`: Array of chat messages to send
+  - `model`: Model ID to use for inference
+  - `params`: Additional model parameters
+  - `retry_delay`: Delay between retries on insufficient balance
+
+- `logging`: Logging configuration
+  - `level`: Log level (DEBUG, INFO, WARNING, ERROR)
+  - `file`: Optional log file path
+
+### Usage
+
+Basic usage:
+
+```bash
+python client.py config.json
+```
+
+Wallet-only mode (just handle payments):
+
+```bash
+python client.py config.json --wallet
+```
+
+Inference-only mode (using existing auth token):
+
+```bash
+python client.py config.json --inference --url "https://inference-url" --key "auth-token"
+```
+
+Override prompt from config:
+
+```bash
+python client.py config.json "Your prompt here"
+```
+
+## Server Components
+
+### Billing Server
+
+The billing server handles WebSocket connections from clients, processes Orchid nanopayments, and maintains client balances. More details coming soon.
+
+### Inference API 
+
+The inference API provides a REST interface for model inference requests, compatible with the OpenAI API format. It integrates with various model providers and handles usage-based billing. More details coming soon.
+
+## Development
+
+Details on setup, contributing, and deployment coming soon.