@recallbricks/agent-cli

Zero-friction onboarding CLI for RecallBricks - AI memory infrastructure.

Get your agent running with memory in 30 seconds:

npx @recallbricks/agent-cli setup rb_live_your_key_here

What is RecallBricks?

RecallBricks provides AI memory infrastructure for your agents. This CLI connects your existing agents to persistent memory with zero configuration.

Quick Start

1. Complete Dashboard Setup

1. Go to dashboard.recallbricks.com/setup
2. Choose your agent type (sales assistant, support agent, etc.)
3. Accept or customize constitutional memories
4. Copy your API key

2. Run the Setup Command

npx @recallbricks/agent-cli setup rb_live_xxxxx

The CLI will:

• Validate your API key
• Detect your project language (TypeScript/JavaScript/Python)
• Ask for your LLM API key (saved for future projects)
• Generate all necessary files
• Install dependencies
• Offer instant testing

3. Start Chatting

# TypeScript
npx ts-node agent.ts

# JavaScript
node agent.js

# Python
python agent.py

Commands

| Command | Description | |---------|-------------| | setup <api_key> | Set up a new agent from dashboard config | | test | Test your agent interactively | | config | Manage saved settings and credentials | | update | Update agent to latest SDK version |

setup

recallbricks setup rb_live_xxxxx [options]

Options:
  -d, --directory <path>   Target directory (default: current)
  -l, --language <lang>    Force language: typescript, javascript, python
  --no-install             Skip dependency installation
  --no-test                Skip test prompt

test

recallbricks test [options]

Options:
  -c, --config <path>      Path to recallbricks-config.json

config

recallbricks config [options]

Options:
  --show                   View saved configuration
  --clear                  Clear all saved credentials
  --set-llm <provider>     Set default LLM (anthropic, openai, google)

update

recallbricks update [options]

Options:
  -d, --directory <path>   Agent directory
  --dry-run                Show changes without applying

Generated Files

After setup, you'll have:

my-agent/
├── agent.ts              # Main agent with memory integration
├── package.json          # Dependencies
├── tsconfig.json         # TypeScript config
├── .env                  # API keys (gitignored)
├── recallbricks-config.json
└── examples/
    ├── 01-simple-chat.ts
    └── 02-memory-usage.ts

Supported LLMs

• Anthropic (Claude) - Default
• OpenAI (GPT-4)
• Google (Gemini)

The CLI auto-detects which LLM you configured in the dashboard and generates the appropriate code.

Example Code

import { chat, rb } from './agent';

// Chat with memory context
const response = await chat("What did we discuss last time?");

// Direct memory operations
await rb.save({
  text: "User prefers dark mode",
  tags: ['preference', 'ui']
});

const memories = await rb.recall({
  query: "user preferences",
  limit: 5
});

Documentation

• Quickstart Guide
• Commands Reference
• Examples
• Troubleshooting

Development

# Install dependencies
npm install

# Run in development
npm run dev -- setup rb_test_xxx

# Build
npm run build

# Run built version
npm start -- setup rb_test_xxx

Links

• RecallBricks Dashboard
• Documentation
• API Reference

License

MIT