// HOW IT WORKS
Four steps to persistent memory
From connection to comprehension in minutes — not days.
Connect
Point your agent at our API with a single config line. Works with OpenClaw, custom agents, and any LLM framework.
Agent Writes Memories
Your agent stores observations, facts, and session notes automatically. We extract structured data and generate embeddings at full resolution.
Semantic Search Retrieves
Three-tier retrieval — handoff notes, recency-boosted recall, and deep semantic search — surfaces exactly what matters for the current context.
Memory Explorer
Browse, search, and edit your agent's memories through a visual UI. Human-readable, human-editable, always in sync.
Agent Graph Explorer
See and explore all of your agent's memories
Every task, skill, document, and preference your agent has learned — visualized as an interactive knowledge graph.
Getting Started
Connect your agent in minutes
No coding required. Hand your agent a prompt and it connects itself. Developers can also use the API directly.
Sign Up & Get Your Connection Package
Create a free account. During onboarding, StateLayer generates a connection package — a set of instructions tailored to your agent platform.
Hand It to Your Agent
Copy the connection package and paste it into your agent's chat, drop it as a file, or text it — however you normally communicate.
Your Agent Remembers
From that point on, your agent writes memories after each session and recalls them at the start of the next. Context carries forward.
Memory Storage
See every memory your agent holds
Browse episodes, facts, decisions, and handoff notes. Search across all memories instantly. Your agent's complete knowledge base at a glance.
// FEATURES
Everything your agent needs to remember
Built for developers who want their agents to learn, retain, and recall — without managing infrastructure.
Hosted & Zero Setup
No databases to provision, no embeddings to manage. Connect your agent and it just works. We handle storage, indexing, and retrieval.
Human-Readable Memories
Every memory is stored as plain text you can read, understand, and edit. No opaque vector blobs — just clean, structured knowledge.
Memory Explorer UI
A visual interface to browse, search, and manage your agent's memories. See what your agent knows, edit what it remembers, delete what it shouldn't.
Multi-Agent Support
One account, multiple agents, fully isolated memory per agent. Scale from a single assistant to an entire fleet without cross-contamination.
Three-Tier Retrieval
Handoff notes for session continuity, recency-boosted search for recent context, and deep semantic recall for long-term knowledge. The right memory at the right time.
Handoff Notes
When a session ends, your agent writes a structured summary of where things stand. The next session picks up exactly where it left off — no context lost.
Automatic Fact Extraction
Our LLM pipeline extracts structured facts, preferences, and decisions from raw conversation. Your agent builds knowledge without manual curation.
// PRICING
Start free. Scale when ready.
All memories stored at full resolution. Memory Explorer included on every plan. Upgrade and search quality improves instantly — zero re-processing.
Get started with persistent memory at no cost.
- 30-day memory retention
- 1 agent
- 1,000 memories
- Standard search quality
- Memory Explorer UI
- REST API access
- Community support
High-resolution search and more agents for growing projects.
- Unlimited retention
- 5 agents
- 50,000 memories
- High-resolution search (1536-dim)
- Memory Explorer UI
- Handoff notes
- Automatic fact extraction
- Email support
For teams running multi-agent systems at scale.
- Unlimited retention
- 20 agents
- 500,000 memories
- High-resolution search (1536-dim)
- Memory Explorer UI
- Full API access
- Team management
- Priority support
- Custom integrations
OPENCLAW + NEMOCLAW INTEGRATION
One command. Persistent memory.
Fully compatible with NemoClaw's sandboxed OpenClaw environment.
Our installer scans your OpenClaw workspace, uploads existing memories, configures semantic search as a tool, and sets up automatic memory sync. Your agent gains persistent, searchable memory without changing how it works.
Works with the tools you already use
OpenClaw Integration
Memory like a lobster — it never forgets
Your agent's memory, supercharged
StateLayer was built for the OpenClaw ecosystem. The installer hooks directly into your workspace — scanning your files, configuring HEARTBEAT.md for scheduled syncs, and adding a memory-search tool so your agent can query its own history. No manual config. No YAML files. Just run the command and go.