@proagentstore/sdk

TypeScript SDK for ProAgentStore — server-powered AI agents.

Install

npm install @proagentstore/sdk

Usage

import { initPro } from '@proagentstore/sdk'

const agent = initPro({
  agentId: 'your-agent-id',
  token: 'your-pags-token',
})

// Chat
const { response } = await agent.chat('Hello!')

// Memory
await agent.memory.set('user-pref', 'preference', 'likes concise answers')
const memories = await agent.memory.list()

// Tasks
await agent.tasks.create('Summarize report', 'Analyze Q3 data')
const tasks = await agent.tasks.list()

// Direct AI inference
const result = await agent.ai.run('@cf/meta/llama-3.2-3b-instruct', {
  messages: [{ role: 'user', content: 'Hello' }]
})

API

initPro(config)

| Parameter | Type | Description | |---|---|---| | agentId | string | Agent or instance ID | | token | string | PAGS session token (from Console → Profile) | | apiBase | string? | API URL (default: https://api.proagentstore.online) |

Returns ProAgentStore

| Method | Description | |---|---| | chat(message) | Send a message, get a response | | messages(limit?) | Get conversation history | | memory.list() | List memory entries | | memory.set(key, type, content) | Store a memory | | memory.delete(key) | Delete a memory | | tasks.list() | List tasks | | tasks.create(title, desc?) | Create a task | | tasks.update(id, status) | Update task status | | ai.run(model, input) | Direct Workers AI inference | | ai.embed(model, text) | Generate embeddings | | subscription.status() | Check subscription |

License

MIT