Add-Ons are isolated, modular services that agents can call, subscribe to or trigger in response to internal or external events. These integrations expand agent functionality beyond core logic, enabling interoperability with APIs, Web3 protocols, hardware controllers and third-party tools.
Each Add-On is registered using a manifest and interacts via the Model Context Protocol (MCP) — FRAKTIΛ’s JSON-RPC-based communication layer for agents.
How Add-Ons Work
Registered via addon.json
manifest
✦ Interfaced through mcp.call()
, mcp.subscribe()
or event triggers.
✦ Run as containerized microservices. (Web2, Web3, robotics, data, etc.)
✦ Assigned permissions at deployment time.
Example Add-On Manifest
Agent Using Add-On
Use Cases
✦ DeFi Bots: price feeds, trade execution, liquidity analysis.
✦ Voice Assistants: weather, scheduling, CRM connections.
✦ Robotic Agents: ROS2, EtherCAT, telemetry data handling.
✦ Monitoring Agents: Discord, X, Telegram listeners with webhook responses.
Security
✦ Scoped API access based on manifest definition.
✦ Enforced rate limits and sandboxing.
✦ Agents require $FRAKT to activate premium Add-Ons.
✦ Governance can restrict or authorize sensitive Add-On scopes.