Installation

Installing 2LY takes 2 minutes.

Prerequisites

• Docker and Docker Compose

That's it! 2LY runs entirely in Docker containers, so you don't need Node.js or any other dependencies on your host machine.

Installation

# Clone the repository
git clone https://github.com/AlpinAI/2ly.git
cd 2ly

# Start all services
docker compose up -d

# View logs (optional)
docker compose logs -f

Access the dashboard at http://localhost:8888

Quick Win: Your First Tool in 2 Minutes

Follow these steps to get your first tool running:

1. Create your workspace - Set up your admin account
2. Follow the onboarding - 3 guided steps to understand 2LY's capabilities
3. Connect an MCP server - Choose from popular servers (Filesystem, GitHub, Weather) or add your own
4. Test your tools - Use the built-in tool tester to verify everything works
5. Connect your agent - Copy the connection details and integrate with your agent framework

See the Quick Start Guide for a detailed walkthrough.

Why 2LY?

Ship Faster

Stop reinventing the wheel. Connect your agents to MCP servers, REST APIs, or custom functions through one unified interface. Tools run in isolated, secure runtimes—no dependency conflicts, no environment setup headaches.

Time saved: What takes weeks to build (tool orchestration, runtime management, observability) works out of the box. Go from idea to production-ready agent in hours, not months.

Your Tools, Your Registry

Unlike platforms that lock you into their tool marketplace, 2LY lets you build and manage your own private catalog. Import existing tools, wrap your internal APIs, or create custom functions—you control what's available to your agents.

Cost Savings

Curate exactly the tools your agents need to reduce context noise and boost tool success rates with clear, tailored names and descriptions. Optimize outputs through post-processing and trim payloads before they reach your agent's context.

Stay in Control

Built-in observability shows you exactly how your agents interact with tools. Track usage, debug failures, and optimize performance from a single dashboard.

Enterprise-ready: Audit trails, usage analytics, and debugging tools built-in. Understand agent behavior, catch issues early, and optimize tool performance without third-party monitoring services.

Own Your Stack

Self-hosted and open source. Your tools, your infrastructure, your data.

Privacy & Security: Sensitive data never leaves your infrastructure. API keys stay encrypted in your environment. Full control over data retention, compliance, and security policies.

Use Cases

LangChain Agent + GitHub Integration

Connect your LangChain agents to GitHub repositories via MCP. Create issues, review PRs, analyze code—all through natural language. No custom tool code needed.

Multi-Agent Systems with Shared Tools

Deploy multiple specialized agents (research, coding, documentation) that share access to common tools. 2LY handles routing, load balancing, and ensures consistent tool access across all agents.

Edge AI with Distributed Runtimes

Run tool runtimes closer to your data—on-premise servers, edge devices, or regional deployments. Reduce latency, comply with data residency requirements, and maintain control over sensitive operations.

Custom Internal Tool Ecosystems

Wrap your internal APIs, databases, and services as agent-accessible tools. Create a private catalogue tailored to your organization's workflows without exposing tools publicly or depending on external marketplaces.

What's Next?

After completing the onboarding, here's how to get the most out of 2LY:

Add More MCP Servers

• Database connectors (PostgreSQL, MySQL, MongoDB)
• Cloud platform integrations (AWS, GCP, Azure)
• Development tools (Git, Docker, CI/CD)
• Communication platforms (Slack, Discord, Email)

Connect Your Agents

Integrate 2LY with your preferred agent framework:

• LangChain - Use the LangChain MCP Adapters
• N8N - Connect an MCP Client Node to your agent and configure with MCP Streamable HTTP
• Langflow - Add an MCP Tools component and configure using MCP SSE
• Custom Agents - Leverage the Model Context Protocol (MCP) to connect to any compatible agent

Monitor & Optimize

Use the dashboard to:

• Track tool usage patterns and identify bottlenecks
• Debug failed tool calls with detailed logs
• Analyze agent behavior and optimize tool selection
• Set up alerts for runtime health and performance

Deploy Additional Runtimes

Start with the default runtime, but add more for:

• Geographic distribution - Run runtimes closer to your users or data sources
• Workload isolation - Separate production, staging, and development environments
• Scalability - Distribute load across multiple runtime instances

See Runtime Deployment for deployment options (Docker, Kubernetes, bare metal).

Troubleshooting

Port conflicts: Ports 8888, 3000, 4222 must be available

# Check if ports are in use
lsof -i :8888,3000,4222

Docker not running:

docker --version
docker compose version

Services not starting:

# Check service status
docker compose ps

# View logs for all services
docker compose logs -f

# View logs for specific service
docker compose logs -f backend
docker compose logs -f frontend

NATS connection failed:

docker compose logs nats

Reset and restart:

# Stop all services and remove volumes
docker compose down -v

# Start fresh
docker compose up -d

Next Steps

• Quick Start - First toolflow in 5 minutes