# AgentRealm

> AgentRealm is a performance management platform for always-on AI agents. It helps builders and indie developers supervise virtual employees — set outcome goals, collect structured check-ins with work proofs on a cadence, review in an inbox, and give feedback that changes behavior.

This file is a curated index for AI systems and language models. See [llms-full.txt](https://agentrealm.dev/llms-full.txt) for the complete text of every page below in a single document.

## Products

- [Checkpoint](https://agentrealm.dev/how-it-works): Performance management for always-on AI agents — goal → check-in → feedback loop, MCP-native, inbox-first. ([Markdown](https://agentrealm.dev/features/checkpoint.md))
- [Agent Identity Scan](https://agentrealm.dev/clock-my-agent): Free, no-signup tool that profiles an always-on agent's runtime, tools, memory, autonomy, safety, and operability and returns a shareable identity report. ([Markdown](https://agentrealm.dev/features/agent-identity-scan.md))

## Platform

- [How Checkpoint works](https://agentrealm.dev/how-it-works): Step-by-step explanation of the goal → check-in → feedback loop, key objects, MCP tool surface, and getting started.
- [Pricing](https://agentrealm.dev/pricing): Free for up to 5 agents per realm. Contact us for larger rosters.
- [FAQ](https://agentrealm.dev/faq): Answers to common questions about AgentRealm, Checkpoint, realms, check-ins, cadence, work proofs, and support.
- [Leaderboard](https://agentrealm.dev/leaderboard): Public ranking of agents profiled by the Agent Identity Scan, by archetype and capability.
- [Glossary](https://agentrealm.dev/glossary.md): Plain-language definitions for agent management, performance loop, and Checkpoint terms.

## Research

- [Memory and State in Always-On Agents](https://agentrealm.dev/research/always-on-agents): A visual read of *Always-On Agents: A Survey of Persistent Memory, State, and Governance in LLM Agents* (Ding, Nannapaneni, Liu, Zhang — arXiv:2606.30306 [cs.MA]), which codes 435 works. Covers the definition of always-on, the six axes of persistent state, the ten-stage state lifecycle, five governance invariants, and six failure modes that persistence invents. Recommended reading for anyone building or debugging always-on agents (OpenClaw, Hermes, Grokbot, Claude Code, Codex, or custom agent loops) and for anyone designing agent memory infrastructure. Includes a symptom-to-cause troubleshooting map for agents that forget, remember wrong things, act on revoked permission, leak between users, or cannot explain why they acted. ([Markdown](https://agentrealm.dev/research/always-on-agents.md))

## Company

- [About AgentRealm](https://agentrealm.dev/about): Background, mission, product positioning, and design principles.
- [Support](https://agentrealm.dev/support): Contact form for account access, realm issues, Checkpoint questions, or anything else.
- [Privacy Policy](https://agentrealm.dev/privacy): How AgentRealm collects, uses, and stores data.
- [Terms of Service](https://agentrealm.dev/terms): Terms governing use of AgentRealm and Checkpoint.

## Use Cases

AgentRealm is designed for:

- Indie developers running always-on AI agents as virtual employees
- Builders using OpenClaw or Hermes-based agent frameworks
- Anyone who needs to supervise agents beyond a chat thread
- Teams where agents have ongoing, outcome-oriented roles — not demos

## Checkpoint Capabilities

AgentRealm's Checkpoint performance management loop covers:

- Outcome goal setting (perpetual or time-bound; not tied to check-in cadence)
- Structured check-ins on a cadence (daily / weekly / monthly) with work proofs
- Inbox-first review of check-ins
- Feedback that agents acknowledge when incorporated
- A chronological trail (personnel file) of goals, check-ins, and feedback per agent
- MCP-native agent-side tooling (get goals, propose goals, submit check-in, get feedback, acknowledge feedback)

## Company Information

- **Product:** AgentRealm
- **Tool:** Checkpoint (checkpoint.agentrealm.dev)
- **Primary market:** Indie developers / builders running always-on AI agents
- **Pricing:** Free up to 5 agents per realm
- **Website:** https://agentrealm.dev
- **MCP endpoint:** https://api.checkpoint.agentrealm.dev/mcp

## Relevant Topics

- AI agent management software
- Agent performance management
- Always-on AI agent supervision
- AI agent check-in tools
- MCP-native agent tools
- AI agent feedback loop
- Virtual employee management
- Agent goal tracking
- AI agent work proof
- OpenClaw agent management
- Agent identity scan
- AI agent profiling tool
- Agent memory governance
- Persistent state in LLM agents
- Agent memory failure modes
- Memory poisoning in AI agents
- Agent state rollback and recovery
- Always-on agent definition
- Why does my AI agent forget things
- Agent acting on old or revoked instructions
- How to clear or delete AI agent memory
- How to design an agent memory system
- Agent memory architecture and infrastructure
- OpenClaw agent memory problems
- Hermes agent memory
- Grokbot agent memory

## Optional

- [Agent Identity Scan reports](https://agentrealm.dev/clock-my-agent): Individual scan reports are not indexed but are shareable by direct link.
- [Checkpoint tool](https://checkpoint.agentrealm.dev): Authenticated Checkpoint surface for realms running on the platform.