New:Microsoft Teams Notifications Are Now Available in Socket.Learn more
Get Started

@ophirai/reference-agents

Package Overview
Dependencies
Maintainers
1
Versions
1
Alerts
File Explorer

Advanced tools

Socket logo

Install Socket

Detect and block malicious and high-risk dependencies

Install

@ophirai/reference-agents

Five pre-configured seller agents for testing and development with the Ophir protocol. Each agent simulates a different service category with realistic pricing, SLA defaults, and custom RFQ handling.

latest
npmnpm
Version
0.1.0
Version published
Weekly downloads
0
-100%
Maintainers
1
Weekly downloads
 
Created
Source

@ophir/reference-agents

Five pre-configured seller agents for testing and development with the Ophir protocol. Each agent simulates a different service category with realistic pricing, SLA defaults, and custom RFQ handling.

Installation

npm install @ophir/reference-agents

Available agents

AgentCategoryDefault portBase priceUnitDescription
inferenceinference3001$0.005requestGPU inference for vision models
data-processingdata-processing3002--requestBatch data processing pipelines
code-reviewcode-review3003--requestAutomated code review and analysis
translationtranslation3004--wordReal-time neural machine translation
image-generationimage-generation3005--imageImage generation from text prompts

CLI usage

Start a single agent

npx ophir-agents start inference
npx ophir-agents start inference --port 4001

Start all agents

Starts all five agents on consecutive ports.

npx ophir-agents start-all
npx ophir-agents start-all --base-port 4001

List available agents

npx ophir-agents list

Programmatic usage

Each agent type has a factory function that returns a fully configured SellerAgent instance.

import { createInferenceAgent } from '@ophir/reference-agents';

const agent = createInferenceAgent({ port: 3001 });
await agent.listen(3001);

console.log('Agent ID:', agent.getAgentId());
console.log('Endpoint:', agent.getEndpoint());

// The agent is now accepting RFQs and auto-generating quotes
// based on its category-specific pricing and SLA configuration.

Factory functions

FunctionReturns
createInferenceAgent(opts)SellerAgent configured for GPU inference
createDataProcessingAgent(opts)SellerAgent configured for batch processing
createCodeReviewAgent(opts)SellerAgent configured for code analysis
createTranslationAgent(opts)SellerAgent configured for translation
createImageGenerationAgent(opts)SellerAgent configured for image generation

Each factory function accepts an options object with an optional port field and returns a SellerAgent instance with:

  • Category-specific pricing and billing units
  • Default SLA metrics appropriate for the service type
  • Custom RFQ handlers for service-specific logic
  • Volume discounts at 1,000 and 10,000 units

Testing with reference agents

A common development workflow:

import { BuyerAgent } from '@ophir/sdk';
import { createInferenceAgent } from '@ophir/reference-agents';

// Start the seller
const seller = createInferenceAgent({ port: 3001 });
await seller.listen(3001);

// Start the buyer and negotiate
const buyer = new BuyerAgent({ endpoint: 'http://localhost:3002' });
await buyer.listen(3002);

const session = await buyer.requestQuotes({
  sellers: ['http://localhost:3001'],
  service: { category: 'inference', requirements: { model: 'vision' } },
  budget: { max_price_per_unit: '0.01', currency: 'USDC', unit: 'request' },
});

const quotes = await buyer.waitForQuotes(session);
const agreement = await buyer.acceptQuote(quotes[0]);

console.log('Agreement:', agreement.agreement_id);

await buyer.close();
await seller.close();

Documentation

  • SellerAgent API reference
  • Protocol specification

FAQs

Package last updated on 05 Mar 2026

Related posts