Add README and SecureBuffer memory locking

This commit is contained in:
Oracle 2026-04-29 19:13:48 +02:00
parent bc1a7a8952
commit 7c6642085f
Signed by: Oracle
SSH key fingerprint: SHA256:x4/RtnjUyuHkdvmwNDsWSfcfF1V5PNr3OpriZqOvCX8
7 changed files with 829 additions and 143 deletions

395
README.md Normal file
View file

@ -0,0 +1,395 @@
# NOMYO Secure Java Chat Client
**OpenAI-compatible secure chat client with end-to-end encryption for NOMYO Inference Endpoints**
🔒 **All prompts and responses are automatically encrypted and decrypted**
🔑 **Uses hybrid encryption (AES-256-GCM + RSA-OAEP with 4096-bit keys)**
🔄 **Drop-in replacement for OpenAI's ChatCompletion API (Java)**
## 🚀 Quick Start
### 0. Try It Now (Demo Credentials)
No account needed — use these public demo credentials to test immediately:
| | |
|---|---|
| **API key** | `NOMYO_AI_E2EE_INFERENCE` |
| **Model** | `Qwen/Qwen3-0.6B` |
> **Note:** The demo endpoint uses a fixed 256-token context window and is intended for evaluation only.
### 1. Installation
via Maven (recommended):
```xml
<dependency>
<groupId>com.nomyo</groupId>
<artifactId>nomyo-java</artifactId>
<version>1.0.0</version>
</dependency>
```
via Gradle:
```groovy
implementation 'com.nomyo:nomyo-java:1.0.0'
```
### 2. Use the client (same API as OpenAI)
```java
import com.nomyo.client.SecureChatCompletion;
import com.nomyo.client.Constants;
import java.util.List;
import java.util.Map;
public class Main {
public static void main(String[] args) {
SecureChatCompletion secureChatCompletion = new SecureChatCompletion(
Constants.DEFAULT_BASE_URL,
"NOMYO_AI_E2EE_INFERENCE"
);
List<Map<String, Object>> messages = List.of(
Map.of("role", "user", "content", "Hello! How are you today?")
);
Map<String, Object> kwargs = Map.of(
"security_tier", "standard",
"temperature", 0.7
);
var response = secureChatCompletion.create(
"Qwen/Qwen3-0.6B",
messages,
kwargs);
System.out.println(response.toString());
}
}
```
## 🔐 Security Features
### Hybrid Encryption
- **Payload encryption**: AES-256-GCM (authenticated encryption)
- **Key exchange**: RSA-OAEP with SHA-256
- **Key size**: 4096-bit RSA keys
- **All communication**: End-to-end encrypted
### Key Management
- **Automatic key generation**: Keys are automatically generated on first use
- **Automatic key loading**: Existing keys are loaded automatically from `client_keys/` directory
- **No manual intervention required**: The library handles key management automatically
- **Keys kept in memory**: Active session keys are stored in memory for performance
- **Optional persistence**: Keys can be saved to `client_keys/` directory for reuse across sessions
- **Password protection**: Optional password encryption for private keys (recommended for production)
- **Secure permissions**: Private keys stored with restricted permissions (600 - owner-only access)
### Secure Memory Protection
### Ephemeral AES Keys
- **Per-request encryption keys**: A unique AES-256 key is generated for each request
- **Automatic rotation**: AES keys are never reused - a fresh key is created for every encryption operation
- **Forward secrecy**: Compromise of one AES key only affects that single request
- **Secure generation**: AES keys are generated using cryptographically secure random number generation
- **Automatic cleanup**: AES keys are zeroed from memory immediately after use
- **Automatic protection**: Plaintext payloads are automatically protected during encryption
- **Prevents memory swapping**: Sensitive data cannot be swapped to disk
- **Guaranteed zeroing**: Memory is zeroed after encryption completes
- **Fallback mechanism**: Graceful degradation if SecureMemory module unavailable
## 🔄 OpenAI Compatibility
The `SecureChatCompletion` class provides **exact API compatibility** with OpenAI's `ChatCompletion.create()` method.
### Supported Parameters
All standard OpenAI parameters are supported:
- `model`: Model identifier
- `messages`: List of message objects (`List<Map<String, Object>>`)
- `temperature`: Sampling temperature (0-2)
- `max_tokens`: Maximum tokens to generate
- `top_p`: Nucleus sampling
- `frequency_penalty`: Frequency penalty
- `presence_penalty`: Presence penalty
- `stop`: Stop sequences
- `n`: Number of completions
- `tools`: Tool definitions
- `tool_choice`: Tool selection strategy
- `user`: User identifier
- And more...
### Response Format
Responses follow the OpenAI format exactly, with an additional `_metadata` field for debugging and security information.
```java
{
"id": "chatcmpl-123",
"object": "chat.completion",
"created": 1234567890,
"model": "Qwen/Qwen3-0.6B",
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": "Hello! I'm doing well, thank you for asking.",
"tool_calls": [...] // if tools were used
},
"finish_reason": "stop"
}
],
"usage": {
"prompt_tokens": 10,
"completion_tokens": 20,
"total_tokens": 30
},
"_metadata": {
"payload_id": "openai-compat-abc123", // Unique identifier for this request
"processed_at": 1765250382, // Timestamp when server processed the request
"is_encrypted": true, // Indicates this response was decrypted
"encryption_algorithm": "hybrid-aes256-rsa4096", // Encryption method used
"response_status": "success" // Status of the decryption/processing
}
}
```
The `_metadata` field contains security-related information about the encrypted communication and is automatically added to all responses.
## 🛠️ Usage Examples
### Basic Chat
```java
import com.nomyo.client.SecureChatCompletion;
import com.nomyo.client.Constants;
import java.util.List;
import java.util.Map;
public class Main {
public static void main(String[] args) {
SecureChatCompletion client = new SecureChatCompletion(
Constants.DEFAULT_BASE_URL,
"NOMYO_AI_E2EE_INFERENCE"
);
List<Map<String, Object>> messages = List.of(
Map.of("role", "system", "content", "You are a helpful assistant."),
Map.of("role", "user", "content", "What is the capital of France?")
);
Map<String, Object> kwargs = Map.of("security_tier", "standard", "temperature", 0.7);
var response = client.create(
"Qwen/Qwen3-0.6B",
messages,
kwargs
);
// Extract content safely
System.out.println(response.get("choices").get(0).get("message").get("content"));
}
}
```
### With Tools
```java
import com.nomyo.client.SecureChatCompletion;
import com.nomyo.client.Constants;
import java.util.List;
import java.util.Map;
public class Main {
public static void main(String[] args) {
SecureChatCompletion client = new SecureChatCompletion(
Constants.DEFAULT_BASE_URL,
"NOMYO_AI_E2EE_INFERENCE"
);
List<Map<String, Object>> messages = List.of(
Map.of("role", "user", "content", "What's the weather in Paris?")
);
Map<String, Object> tools = Map.of(
"type", "function",
"function", Map.of(
"name", "get_weather",
"description", "Get weather information",
"parameters", Map.of(
"type", "object",
"properties", Map.of(
"location", Map.of("type", "string")
),
"required", List.of("location")
)
)
);
Map<String, Object> kwargs = Map.of("security_tier", "standard", "temperature", 0.7);
var response = client.create(
"Qwen/Qwen3-0.6B",
messages,
kwargs
);
System.out.println(response.get("choices").get(0).get("message").get("content"));
}
}
```
## 📦 Dependencies
- **Maven Coordinates**: `com.nomyo:nomyo-java`
- **Java Version**: Java 11+ (Required for HTTP Client and crypto primitives)
- **JSON Processing**: Jackson Databind (for Map/JSON handling)
- **HTTP Client**: Apache HttpClient or Java NIO (Included in core)
- **Crypto**: BouncyCastle or JDK Crypto (Included in core)
## 🔧 Configuration
### Custom Base URL
```java
SecureChatCompletion client = new SecureChatCompletion("https://NOMYO-Pro-Router:12434", "YOUR_API_KEY");
```
### API Key Authentication
```java
// Initialize with API key (recommended for production)
SecureChatCompletion client = new SecureChatCompletion(
"https://api.nomyo.ai",
"your-api-key-here"
);
// Or pass API key in the create() method if supported by extension
// Map<String, Object> kwargs = Map.of("security_tier", "standard", "api_key", "your-api-key-here");
```
### Secure Memory Configuration
The library enables secure memory protection by default.
```java
// Enable secure memory protection (default)
SecureChatCompletion client = new SecureChatCompletion("https://api.nomyo.ai", "YOUR_API_KEY");
// Disable secure memory (not recommended, for testing only)
// Requires passing specific config flag if available in version
SecureChatCompletion client = new SecureChatCompletion("https://api.nomyo.ai", "YOUR_API_KEY");
```
### Key Management
Keys are automatically generated on first use.
#### Generate Keys Manually
```java
import com.nomyo.client.SecureCompletionClient;
public class KeyManager {
public static void main(String[] args) {
SecureCompletionClient client = new SecureCompletionClient();
client.generateKeys(true, "client_keys", "your-password");
}
}
```
#### Load Existing Keys
```java
import com.nomyo.client.SecureCompletionClient;
public class KeyManager {
public static void main(String[] args) {
SecureCompletionClient client = new SecureCompletionClient();
client.loadKeys("client_keys/private_key.pem", "client_keys/public_key.pem", "your-password");
}
}
```
## 📚 API Reference
### SecureChatCompletion
#### Constructor
```java
SecureChatCompletion(
String base_url,
String api_key
)
```
**Parameters:**
- `base_url`: Base URL of the NOMYO Router (must use HTTPS for production)
- `api_key`: Optional API key for bearer authentication
#### Methods
- `create(String model, List<Map<String, Object>> messages, Map<String, Object> kwargs)`: Create a chat completion
### SecureCompletionClient
#### Constructor
```java
SecureCompletionClient(
String router_url
)
```
#### Methods
- `generateKeys(boolean saveToFile, String keyDir, String password)`: Generate RSA key pair
- `loadKeys(String private_key_path, String public_key_path, String password)`: Load keys from files
- `fetchServerPublicKey()`: Fetch server's public key
- `encryptPayload(Map<String, Object> payload)`: Encrypt a payload
- `decryptResponse(Map<String, Object> encrypted_response, String payload_id)`: Decrypt a response
- `sendSecureRequest(Map<String, Object> payload, String payload_id)`: Send encrypted request and receive decrypted response
## 📝 Notes
### Security Best Practices
- Always use password protection for private keys in production
- Keep private keys secure (permissions set to 600)
- Never share your private key
- Verify server's public key fingerprint before first use
### Performance
- Key generation takes ~1-2 seconds (one-time operation)
- Encryption/decryption adds minimal overhead (~10-20ms per request)
### Compatibility
- Works with any OpenAI-compatible code
- No changes needed to existing OpenAI client code
- Simply replace `openai.ChatCompletion.create()` with `SecureChatCompletion.create()`
## 🤝 Contributing
Contributions are welcome! Please open issues or pull requests on the project repository.
## 📄 License
See LICENSE file for licensing information.
## 📞 Support
For questions or issues, please refer to the project documentation or open an issue.