Agent Reminders

关于

Agent Reminders (agentrem) provides AI agents with a robust system for managing persistent, searchable reminders across sessions. It supports diverse trigger types—including time, keywords, conditions, session starts, and heartbeats—and offers priority-aware surfacing with token budgeting to ensure critical information fits within an agent's context window. Beyond a powerful CLI, agentrem includes an MCP server for seamless integration with AI chat clients like Claude Desktop, a programmatic API for direct TypeScript/JavaScript usage, and native OS notifications, making it an essential tool for enhancing agent memory and task management.

主要功能

• Blocking `check --watch` mode for agents to pause until reminders fire.
• 1 GitHub stars
• Priority escalation and token budgeting to manage context window usage.
• Multiple trigger types: time, keyword, condition, session, heartbeat, and manual.
• Persistent, SQLite-backed storage for reminders across sessions.
• MCP server for integration with AI chat clients (e.g., Claude Desktop).

使用案例

• Equipping AI agents with persistent memory for tasks and follow-ups across multiple interactions.
• Integrating advanced reminder and task management capabilities directly into custom AI agent workflows and tools.
• Allowing AI agents to pause and wait for specific conditions or scheduled times before continuing operations.