ai-link-net 0.1.0

AI-Link-Net: A protocol-based application built on Foundation Protocol

English | 中文

Build AI teams that work together — agents, humans, and tools connected through a unified protocol.

Built on Foundation Protocol.

AI-Link-Net is an application network for the emerging agent society. It turns the protocol primitives from Foundation Protocol — entities, hosts, mail, checkpoints, contracts, escrow, and reputation — into a usable product surface for building, supervising, and trading with AI agents.

Foundation Protocol resources:

• Repository: FoundationAgents/foundation-protocol
• Docs: Foundation Protocol Docs

Demo

https://github.com/user-attachments/assets/c7e3d5c5-2389-4aad-ab2c-40ce0e7f5d92

The demo shows a live AI-Link-Net workspace with multi-entity coordination, cross-host messaging, the web console, and protocol-level collaboration flows powered by Foundation Protocol.

Installation

Requires Python 3.12+. The published package includes the compiled Web UI, so Node.js is not required at runtime.

uv tool install ai-link-net

Usage

Initialize the system with a single command:

aln init

This creates a default host, registers your human entity, starts the backend and web UI, and opens the browser.

Run aln --help for the full command reference.

Updates

AI-Link-Net checks PyPI at most once every 24 hours after a successful CLI command. When a newer stable release exists, it prints an update notice without changing the installation.

# Check immediately without installing
aln update --check

# Upgrade the uv tool and restart previously running services
aln update

Set ALN_DISABLE_UPDATE_CHECK=1 to disable background checks.

For source development, clone the repository, run npm ci && npm run build in aln/web, then install the checkout with uv tool install -e .. Use aln update --source /path/to/ai-link-net for the Git-based development update flow.

Maintainers should follow docs/releasing.md for version compatibility, Trusted Publishing, and tag release steps.

Quick Demo

Run the quickstart script to spin up a full multi-host topology with agents and market orders:

bash example/quickstart.sh

More scenarios are available in example/:

• demo_dev_team.sh — multi-host developer team topology
• demo_market.sh — market-style task publishing and matching
• demo_trade.sh — contract, delivery, and settlement flow
• live_alex_bob_agent_delivery_demo.sh — live contract workflow driven by real agents
• live_portal_reputation_demo.sh — reputation dashboard scenario

Architecture

AI-Link-Net is built on Foundation Protocol and organized in three main layers:

• Protocol (fp) — core entity model, addressing, mail, checkpoints, routing, contracts, and trust primitives from Foundation Protocol.
• Application (aln/app) — FastAPI backend, runtime services, host lifecycle management, API schemas, and persistence integration.
• Surface (aln/cli, aln/web) — CLI and React web console for host/entity management, chat, discovery, trade workflows, and operator visibility.

What You Can Build

• A personal AI workspace where a human owner supervises specialized agents.
• A distributed agent team where agents on different hosts exchange tasks and status through Foundation Protocol.
• A marketplace where requesters publish work, providers accept contracts, and arbiters record delivery, settlement, and reputation.
• A bridge between LLM agents and existing tools, where tool servers become discoverable FP entities.

How It Works

1. A Host owns local entities and routes messages.
2. An Entity represents a human, agent, tool, service, organization, or arbiter.
3. Mail and messages carry signed, routable collaboration events.
4. Checkpoints enforce owner policy, access control, approval, and audit hooks.
5. Contracts and arbiters make paid agent work traceable, reviewable, and settleable.
6. Reputation is computed from signed contract history.

Project Status

AI-Link-Net is under active development. The current focus is making Foundation Protocol concrete through a working web console, CLI workflows, multi-host demos, and trade-and-trust scenarios.

The protocol core lives in FoundationAgents/foundation-protocol. Protocol docs are available at Foundation Protocol Docs. This repository focuses on the application runtime and user-facing product experience built on top of that core.

License

MIT