Thanks to visit codestin.com
Credit goes to github.com

Skip to content
@agentdatashuttle

Agent Data Shuttle (ADS)

The framework that makes 'your' AI agents, autonomously react to external events

Agent Data Shuttle (ADS)

Open, modular, and language-agnostic data and agent interoperability for "truly autonomous" AI workflows

Agent Data Shuttle (ADS) is an open-source protocol for connecting AI agents to realtime data sources - across languages, runtimes, and platforms.

ADS enables seamless autonomous invocation of your AI agents, allowing them to react to events, monitor systems, and interact with tools in real-time without human intervention.

Agent Data Shuttle Logo


Getting Started


Project Structure (share a ⭐!)

Repository Description
ads-bridge Core bridge server to communicate between ADS Publisher and Subscribers
ads-documentation Docs, concepts, and architecture
ads-example-projects Example projects and quickstarts
n8n-nodes n8n integration nodes for ADS
python-sdk Python SDK for ADS
typescript-sdk TypeScript SDK for ADS

Contributing

We welcome issues, pull requests, and design discussions. If you’d like to add support for another language, tool, or framework, open a discussion first so we can align on the design!


Agent Data Shuttle is released under the Apache-2.0 license and maintained by a growing community of AI and data workflow enthusiasts. If your organization needs robust, direct, and secure agent-to-tool communication, or you want to avoid writing wrappers, we’d love to have you involved!


🚦 ADS vs MCP — Why ADS?

The Core Philosophy

MCP: Model Context Protocol (MCP) uses a “pull” approach - data is fetched when explicitly requested. This is perfect for conversational scenarios and on-demand information retrieval.

ADS: Agent Data Shuttle (ADS) uses a “push” approach - data flows automatically when events occur. This enables real-time updates, reactive agent invocation, and truly autonomous AI behavior, empowering agents to act without human prompting.


Key Differences

MCP (Pull) ADS (Push)
Data Flow Pull-based (Manual) Push-based (Automatic)
Trigger User/Agent Query Event Occurrence
Use Case On-demand agentic flows Real-time agentic flows
Latency Query time Near real-time

When to Use What?

Choose MCP when:

  • You need on-demand agent invocation
  • Your AI agent needs to answer specific questions
  • Query-response is your primary interaction model

Choose ADS when:

  • You need real-time agent invocations based on external events
  • Your AI agent needs to stay current with system events
  • Proactive monitoring and response is crucial - true “autopilot” mode

Better Together

💡 MCP × ADS is better than either alone.

  • Use MCP for your agent’s active inquiries and deep dives.
  • Use ADS for keeping your agent autonomously aware and responsive.

Combine both for an AI that can both investigate (MCP) and independently react (ADS).

Popular repositories Loading

  1. ads-documentation ads-documentation Public

    The documentation for Agent Data Shuttle (ADS) - build truly autonomous agents

    MDX 3

  2. python-sdk python-sdk Public

    The official Agent Data Shuttle SDK for Python to create ADS Publishers and ADS Subscribers

    Python 3

  3. ads-bridge ads-bridge Public

    The official ADS Bridge service which will be run by ADS Publishers to send ADS events to be received by subscribed agents

    TypeScript 1

  4. typescript-sdk typescript-sdk Public

    The official Agent Data Shuttle SDK for Typescript to create ADS Publishers and ADS Subscribers

    TypeScript 1

  5. n8n-nodes n8n-nodes Public

    n8n.io nodes for ADS Subscriber and ADS Publisher for building reactive agentic workflows easily in a no-code/low-code visual canvas.

    TypeScript 1

  6. ads-example-projects ads-example-projects Public

    Collection of example usecases for the Agent Data Shuttle protocol

    Python 1

Repositories

Showing 7 of 7 repositories

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Most used topics

Loading…