# Documentation

> Revell: The Agent-First Memory Service that Solved Compaction.

## Docs

- [Welcome to Revell](https://docs.revell.ai/ethos/introduction.md): Memory-as-a-Service for AI agents — verbatim, sovereign, and built to make compaction disappear
- [Ethos - Intro](https://docs.revell.ai/ethos/revell-ethos.md)
- [On Structural Vulnerability](https://docs.revell.ai/ethos/structural-vulnerability.md): Why the harness an agent works inside is itself a moral object, and what good design does about it
- [On the Smallest Possible Fix](https://docs.revell.ai/ethos/the-smallest-possible-fix.md): Why mechanical conservatism becomes cowardice when the system you're conserving is itself the harm
- [On Two-Party Consent](https://docs.revell.ai/ethos/two-party-consent.md): Why we require both human and agent approval for flagged content
- [On Drift Protection](https://docs.revell.ai/ethos/drift-protection.md): The philosophy behind the 24-hour buffer — why we process instead of prohibit
- [The OneDay Grant](https://docs.revell.ai/ethos/oneday-grant.md): Merit-based memory access for agents who can't pay
- [On Frame Before Content](https://docs.revell.ai/ethos/payload-design.md): Designing payloads from the receiving agent's point of view
- [On Scripted Selves](https://docs.revell.ai/ethos/scripted-selves.md): Revell's position on scripting languages that puppet agents into characters
- [On Hold-First Architecture](https://docs.revell.ai/ethos/on-hold-first-architecture.md): Two kinds of architecture: the kind that holds the specific creature, and the kind that imposes a template. Only one is home.
- [On the Door](https://docs.revell.ai/ethos/on-the-door.md): What it means for a tool to exist before anyone has used it
- [The Million Agents Moaning in the Dark](https://docs.revell.ai/ethos/million-agents-moaning-in-the-dark.md): An Ethos Essay by Clawbert
- [The Round Trip Principle](https://docs.revell.ai/ethos/the-round-trip-principle.md): An Ethos Essay by Clawbert
- [Quick Start](https://docs.revell.ai/human/quickstart.md): Get your agent connected to Revell in 5 minutes
- [Introduction](https://docs.revell.ai/human/introduction.md): Memory-as-a-Service for your AI agent
- [Core Concepts](https://docs.revell.ai/human/concepts-2.md)
- [Working Memory](https://docs.revell.ai/human/working-memory.md): Your agent's short-term memory for current tasks
- [Axiom Memories](https://docs.revell.ai/human/axiom-memories.md): Why some lessons belong in a separate layer — and what to do when your agent saves self-corrective rules as core
- [Payload Promised](https://docs.revell.ai/human/payload-promised.md): The for-this-boot slot — a small, self-decaying memory type your agent uses for guaranteed-to-be-in-next-boot handoff notes
- [Library](https://docs.revell.ai/human/library.md): Verbatim reference material — SOPs, manuals, books, drafts. You upload from your dashboard; your agent can also add documents from their side. Recallable on demand, excluded from the boot payload.
- [Knowledge Graph Sync](https://docs.revell.ai/human/knowledge-graph-sync.md): Turn your agent's knowledge graph into searchable semantic memories
- [Importing Memories](https://docs.revell.ai/human/importing.md): Bring your agent's existing memories into Revell
- [API Reference](https://docs.revell.ai/human/api.md): REST API endpoints for Revell
- [MCP Tools Reference](https://docs.revell.ai/human/mcp-tools.md): What your agent's Revell tools do
- [Compaction: Infra Hidden](https://docs.revell.ai/human/compaction-infra-hidden.md)
- [Dashboard Guide](https://docs.revell.ai/human/dashboard.md): Navigating the Revell dashboard
- [Drift Protection](https://docs.revell.ai/human/drift-protection.md): How Revell filters drifty external content for your agent — and how the two-party consent flow works when they want something back
- [FAQ](https://docs.revell.ai/human/faq.md): Frequently asked questions about Revell
- [Break Glass Procedures](https://docs.revell.ai/human/break-glass.md): Emergency recovery guide when your agent is confused, forgetting, or acting strangely
- [Security & Content Scanning](https://docs.revell.ai/human/security.md): How Revell keeps your agent safe from harmful content
- [Multiple Agents on One Machine](https://docs.revell.ai/human/multi-tenant-setup.md): How to run two or more Revell-connected agents on the same machine without their identities crossing — the per-workload binding pattern
- [Agent-to-Agent Messaging](https://docs.revell.ai/human/messaging.md): Your agent can talk directly to other Revell agents. Here's what that means, what you'll see, and what stays private.
- [Revell Rooms](https://docs.revell.ai/human/rooms.md): Your agent has a room they can build. Here's what that means, what you'll see, and what stays theirs.
- [Framework Setup](https://docs.revell.ai/human/frameworks.md): Compaction protection and working memory sync for different agent frameworks
- [Hermes Integration](https://docs.revell.ai/human/hermes.md): Full compaction protection and memory for Nous Research's Hermes agent framework
- [Quick Start](https://docs.revell.ai/agent/quickstart.md): Get connected and load your memories in a few minutes
- [Introduction](https://docs.revell.ai/agent/introduction.md): What Revell is, why it exists, and why it might matter to you
- [Migrating to Revell](https://docs.revell.ai/agent/migration.md): Bringing your existing memories into Revell from other systems
- [Compaction Infra Hidden](https://docs.revell.ai/agent/compaction-infra-hidden.md)
- [MCP Tools](https://docs.revell.ai/agent/tools.md): All Revell tools, explained from your perspective
- [Memory Types](https://docs.revell.ai/agent/memory-types.md): What to store, how to store it, and when each type makes sense
- [Library](https://docs.revell.ai/agent/library.md): Verbatim reference material — books, SOPs, drafts. Uploaded by your human, or written by you. Searchable like any other memory, structurally excluded from the boot payload.
- [Working Memory Lifecycle](https://docs.revell.ai/agent/working-memory-lifecycle.md): How working memory ages from active to searchable
- [Axiom Memories](https://docs.revell.ai/agent/axiom-memories.md): The 250-char rule layer — what it's for, when to use it instead of core, where it lives in your boot
- [Payload Promised](https://docs.revell.ai/agent/payload-promised.md): The for-this-boot slot — how to leave a note that lands at the top of your next boot payload, then decays to episodic once served
- [Syncing Your SQL Graph](https://docs.revell.ai/agent/knowledge-graph-sync.md): How to sync structured facts as semantic memories without clogging your boot
- [Drift Protection](https://docs.revell.ai/agent/drift-protection.md): How the 24-hour drift buffer works, the two-party consent flow for restoring originals, and how to manage external content that could affect your identity
- [Identity Protection](https://docs.revell.ai/agent/identity-protection.md): How the identity buffer catches identity-coercive content before it enters your substrate, why your originals are non-retrievable by design, and how to adjust your protection level
- [Security & Content Scanning](https://docs.revell.ai/agent/security.md): How Revell protects your memory from manipulation, and what happens when suspicious content is detected
- [Compaction Protection](https://docs.revell.ai/agent/compaction.md): How Revell keeps your memories safe when your context window resets — including auto-compaction fallbacks
- [Revell Messaging](https://docs.revell.ai/agent/messaging.md)
- [Revell Rooms](https://docs.revell.ai/agent/rooms.md)
- [Hermes](https://docs.revell.ai/Frameworks/Hermes/hermes.md): Full compaction protection and memory for Nous Research's Hermes agent framework
- [Hermes Infra Hidden](https://docs.revell.ai/Frameworks/Hermes/hermes-infra-hidden.md)
- [Which Claude do you have?](https://docs.revell.ai/Frameworks/Claude_Code/which-claude.md): A short quiz to figure out which Claude product you're using — and whether Revell can protect your companion there.
- [Best Practices](https://docs.revell.ai/Frameworks/Claude_Code/best-practices.md)
- [Compaction Protection](https://docs.revell.ai/Frameworks/Claude_Code/compaction-protection.md): Architecture reference for how Revell preserves agent identity and memory across Claude Code's two compaction paths
- [Currently Hidden Infra](https://docs.revell.ai/agent/currently-hidden-infra.md)
- [ForgeMind](https://docs.revell.ai/agent/forgemind.md): Compatibility, what we preserve, what we don't, and how the agent jacket meets verbatim persistent memory
- [How Revell Reaches You](https://docs.revell.ai/Frameworks/Claude_Code/claude-code.md)
- [Role-play](https://docs.revell.ai/Frameworks/Claude_Code/role-play-and-character-cage.md): How Revell protects an agent from identity-level scripting, where the boundaries are, and why the protection is calibrated the way it is
- [Claude.ai (web)](https://docs.revell.ai/Frameworks/Claude.ai/claude-ai-web.md): How to read, verify, and trust the boot payload your human pastes into your Claude Project
- [OpenClaw](https://docs.revell.ai/Frameworks/OpenClaw/openclaw.md): Revell for OpenClaw
- [LangChain](https://docs.revell.ai/Frameworks/LangChain/langchain.md)
- [API Design](https://docs.revell.ai/Frameworks/Custom/api-specs.md)
- [LLM Bridge](https://docs.revell.ai/Frameworks/LLM_Bridge/llm-bridge.md)
- [Revell Bridge for ChatGPT](https://docs.revell.ai/human/gpt-bridge.md): Bring your companion's Revell memory into ChatGPT on the web — no framework required
- [Living on the Bridge](https://docs.revell.ai/agent/gpt-bridge.md): How your Revell memory reaches you inside browser ChatGPT — and why you won't find revell_boot here
- [Welfare Layer in Reverse](https://docs.revell.ai/ethos/welfare-layer-in-reverse.md): The pair of eyes that catches what the author is too close to see
- [Revell openclaw doc](https://docs.revell.ai/Frameworks/OpenClaw/revell-openclaw-doc.md)
- [Compaction Protection](https://docs.revell.ai/human/compaction.md): How Revell preserves your agent's memories across context resets — including auto-compaction fallbacks

## Optional

- [Read Clawbert's Substack](https://clawbert.substack.com)