https://github.com/layr-labs/eigen-agentkit
Eigen AgentKit
https://github.com/layr-labs/eigen-agentkit
Last synced: 19 days ago
JSON representation
Eigen AgentKit
- Host: GitHub
- URL: https://github.com/layr-labs/eigen-agentkit
- Owner: Layr-Labs
- License: mit
- Created: 2025-02-08T22:57:06.000Z (4 months ago)
- Default Branch: master
- Last Pushed: 2025-02-15T22:36:57.000Z (4 months ago)
- Last Synced: 2025-05-06T20:43:18.999Z (26 days ago)
- Language: TypeScript
- Size: 310 KB
- Stars: 7
- Watchers: 3
- Forks: 8
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# @layr-labs/agentkit
A modular and extensible agent framework that provides verifiable AI capabilities and data availability logging. It integrates multiple tools under one roof:
- [Opacity](https://github.com/Layr-Labs/eigen-agentkit/tree/master/packages/adapter-opacity) for verifiable inference
- [EigenDA](https://github.com/Layr-Labs/eigen-agentkit/tree/master/packages/adapter-eigenda) for data availability logging
- [WitnessChain](https://github.com/Layr-Labs/eigen-agentkit/tree/master/packages/adapter-witnesschain) for Real-world actuation and observation
- Reclaim for verifiable API calls (TBD)
- Formation for verifiable code execution (TBD)
- Silence for secret publishing (TBD)
## Features
- 🔒 Verifiable AI inference with zkTLS proofs
- 📝 Data availability logging with EigenDA
- 📍 Real-world actuation and observation with Witnesschain InfinityWatch
- 🔑 Verifiable API calls and external data integration
- ⚡ Composable with existing AI frameworks (LangChain, etc.)
- 🛠️ Modular adapter system for extensibility
## Project Structure
```
packages/
├── core/ - Core interfaces and types
├── adapter-opacity/ - Opacity adapter for verifiable inference
├── adapter-eigenda/ - EigenDA adapter for data availability logging
├── adapter-witnesschain/ - Witnesschain adapter for InfinityWatch
├── adapter-reclaim/ - Reclaim adapter for API calls (TBD)
├── adapter-formation/ - Formation adapter for code execution (TBD)
├── adapter-silence/ - Silence adapter for secret publishing (TBD)
└── agent/ - Main agent implementation (TBD)
```
## Getting Started
### Prerequisites
- Node.js >= 18
- pnpm >= 8.15.1
### Installation
```bash
# Install dependencies
pnpm install
# Build all packages
pnpm build
# Run tests
pnpm test
```
### Environment Variables
Create a `.env` file in the root directory with the following variables:
```env
# Opacity Configuration
OPACITY_API_KEY=your_api_key
OPACITY_TEAM_ID=f309ac8ae8a9a14a7e62cd1a521b1c5f
OPACITY_TEAM_NAME=eigen-test
OPACITY_PROVER_URL=https://opacity-ai-zktls-demo.vercel.app
# EigenDA Configuration
EIGENDA_PRIVATE_KEY=your_private_key
EIGENDA_API_URL=https://test-agent-proxy-api.eigenda.xyz
EIGENDA_BASE_RPC_URL=https://mainnet.base.org
EIGENDA_CREDITS_CONTRACT=0x0CC001F1bDe9cd129092d4d24D935DB985Ce42A9
# Witnesschain Configuration
WITNESSCHAIN_API_KEY=your_api_key
WITNESSCHAIN_API_URL=https://api.witnesschain.com
WITNESSCHAIN_PRIVATE_KEY=your_private_key
```
## Usage
### Verifiable Inference with Opacity
```typescript
import { OpacityAdapter } from '@layr-labs/agentkit-opacity';
// Initialize adapter
const opacityAdapter = new OpacityAdapter({
teamId: process.env.OPACITY_TEAM_ID!,
teamName: process.env.OPACITY_TEAM_NAME!,
apiKey: process.env.OPACITY_API_KEY!,
opacityProverUrl: process.env.OPACITY_PROVER_URL!,
});
// Generate text with proof
const result = await opacityAdapter.generateText('Your prompt here');
console.log('Generated text:', result.content);
console.log('Proof:', result.proof);
// Verify the proof
const isValid = await opacityAdapter.verifyProof(result.proof);
console.log('Proof is valid:', isValid);
```
### Data Availability Logging with EigenDA
```typescript
import { EigenDAAdapter } from '@layr-labs/agentkit-eigenda';
// Initialize adapter
const eigenDAAdapter = new EigenDAAdapter({
privateKey: process.env.EIGENDA_PRIVATE_KEY!,
// Optional
apiUrl: process.env.EIGENDA_API_URL,
rpcUrl: process.env.EIGENDA_BASE_RPC_URL,
creditsContractAddress: process.env.EIGENDA_CREDITS_CONTRACT,
});
// Initialize the adapter (required before logging)
await eigenDAAdapter.initialize();
// Log messages with different levels
await eigenDAAdapter.info('Application started', { version: '1.0.0' });
await eigenDAAdapter.warn('Resource usage high', { cpu: 90, memory: 85 });
await eigenDAAdapter.error('Failed to connect', { service: 'database' });
await eigenDAAdapter.debug('Processing request', { requestId: '123' });
// Check if a log is available
const logEntry = await eigenDAAdapter.info('Test message');
const isAvailable = await eigenDAAdapter.checkAvailability(logEntry.status);
console.log('Log is available:', isAvailable);
// Retrieve a log entry
const retrievedLog = await eigenDAAdapter.getLogEntry(logEntry.id);
console.log('Retrieved log:', retrievedLog);
// Clean up when done
await eigenDAAdapter.shutdown();
```
### Real-world Actuation and Observation with Witnesschain InfinityWatch
The **Witnesschain Adapter** is a TypeScript-based utility that enables an agent to request tasks and observations in the real world. It works in conjunction with the **InfinityWatch app**, which acts as a portal to the physical world.
#### **Usage**
```typescript
import { WitnesschainAdapter } from '@layr-labs/agentkit-witnesschain';
// Initialize adapter
const witnesschain = new WitnesschainAdapter();
// Authenticate with Ethereum-compatible wallet
const isLoggedIn = await witnesschain.login();
console.log("Login successful:", isLoggedIn);
// Create a new campaign
const campaign = await witnesschain.createCampaign({
campaign: "Urban Pollution Check",
description: "Collect images of high-smog areas for air quality analysis.",
latitude: 37.7749,
longitude: -122.4194,
radius: 50, // Radius in kilometers
reward_per_task: 5,
total_rewards: 100,
starts_at: new Date().toISOString(),
ends_at: new Date(Date.now() + 7 * 24 * 60 * 60 * 1000).toISOString(),
is_active: true
});
console.log("Campaign Created:", campaign);
// Fetch geoverified observations
const observations = await witnesschain.getCampaignPhotos("Urban Pollution Check", null);
console.log("Received Observations:", observations);
// Classify and accept photos
const imagePaths = ["smog1.jpg", "smog2.jpg"];
const task = "Identify high-smog areas";
const classification = await witnesschain.classifyPhotos(imagePaths, task);
console.log("Classification Result:", classification);
if (classification.success) {
await witnesschain.acceptPhoto("smog1.jpg");
}
```
## Development
```bash
# Start development mode
pnpm dev
# Lint code
pnpm lint
# Format code
pnpm format
```
## Contributing
Contributions are welcome! Please read our contributing guidelines (coming soon) for details on our code of conduct and the process for submitting pull requests.
## Contributors
Thanks to all our contributors who make AgentKit possible!
## License
This project is licensed under the MIT License - see the LICENSE file for details.