Stop Re-Teaching Agents

The unified agent memory layer.

Segment agent memory by lifecycle: session context, semantic truths, episodic logs, and procedural workflows. Recall what matters and generate native configs for every tool.

$pnpm add -g @memories.sh/cli

Live StateLocal

Ask your agent to pick up where you left off...

“Continue the auth refactor with the same error-handling rules.”

Recent Context

Use Tailwind for all UI componentsSemantic

Checkpoint saved before auto compactionEpisodic

Release checklist skill promoted after successful runsProcedural

Segment, recall, and generate in one flow

Works with

Memory Segmentation

One memory system. Four distinct lanes.

This is the core model behind memories.sh. Session, semantic, episodic, and procedural memory are stored separately so agents keep continuity without blending durable truths with transient chat history.

Session

Active working context for the current task.

Stored: Session events, checkpoints, boundary snapshots.

Recall: Used first for continuity in ongoing work.

Semantic

Durable truths and stable project knowledge.

Stored: Rules, facts, decisions, `memory.md` in OpenClaw mode.

Recall: Always-on grounding for consistent behavior.

Episodic

Timeline fidelity for what happened and when.

Stored: Daily logs + raw session snapshots.

Recall: Chronology and context history on demand.

Procedural

Repeatable workflow memory and runbooks.

Stored: Skill-like workflows and promotion signals.

Recall: Prioritized when intent matches a workflow.

Compaction Triggers

Count Trigger

Near token/turn budget boundaries

Time Trigger

Session inactivity compaction windows

Event Trigger

Reset, handoff, or task-complete boundaries

Read Segmentation Docs Explore Session Commands

How It Works

Three commands. That's it.

Capture session checkpoints, recall layered context, and generate native configs for every agent tool from one CLI.

Segmented Lifecycle

Session

Active working context with checkpoints and boundary snapshots.

Long-term

Semantic truths plus episodic daily logs and raw snapshots.

Procedural

Reusable workflows recalled when task intent matches.

01 - Command Line

>pnpm add -g @memories.sh/cli

Quickstart Guide

Store, recall, generate

Session lifecycle + long-term semantic, episodic, and procedural memory in local SQLite.

Learn more

Terminal

~/project

Stored in local semantic memory

CLIStore, recall, generate

Core Features

Built for segmented memory

Lifecycle-aware memory for real agent work: session continuity, compaction safety, and long-term retrieval that stays coherent.

Segmented Memory Architecture

Separate session working context from long-term semantic, episodic, and procedural memory so agents stop mixing timelines with durable truths.

4 memory lanes

Learn More →

Compaction-Safe Checkpoints

Write checkpoints and snapshots before context boundaries so important state survives reset, inactivity, and token pressure.

Boundary-safe

Learn More →

Session + Long-Term Recall

Recall merges always-on rules with relevant session and long-term memories, so answers stay consistent across long tasks.

Layered context

Learn More →

Tool-Native Output

Generate native configs for Cursor, Claude Code, Copilot, Windsurf, Cline, Roo, Gemini, and more.

One command

Learn More →

Deterministic OpenClaw File Mode

Run bootstrap, flush, and snapshot against memory.md, daily logs, and raw snapshots for git-friendly lifecycle continuity.

bootstrap/flush/snapshot

Learn More →

Portable State

Export everything to JSON or YAML anytime. Switch tools or leave entirely. Your data is yours.

No lock-in

Learn More →

TypeScript SDK

Cloud-native memory for AI apps and dev teams.

The SDK connects your AI apps to memories.sh cloud—giving every prompt the right rules and context. memoriesMiddleware() is all it takes.

Start in Dashboard → AI SDK Projects to create your API key and project isolation (`tenantId`). Then add middleware and ship. The CLI and MCP handle local-first storage and agent access, while the SDK handles your app runtime.

Read the SDK docs

SDK@memories.sh/ai-sdk

memoriesMiddleware() connects any model to cloud memory. Pair with AI SDK Projects in the dashboard.

MCPmemories serve

Hosted MCP exposes tenant-routed core memory tools; local memories serve adds lifecycle, reminder, and streaming tools.

CLInpx memories

Local SQLite, embeddings, 13+ tool configs, offline-first. The foundation.

↑ built on ↑

API Contract

Production API that actually ships fast.

One predictable SDK surface for retrieval, writes, graph rollout, and diagnostics. Designed for strict clients and fast incident debugging. `tenantId` is the database boundary, `userId` is end-user scope, and `projectId` is an optional repo context filter. Default SDK transport is HTTP API; MCP transport is optional.

POST/api/sdk/v1/context/get

Pull rules + memories with scoped routing and graph-aware retrieval strategy controls.

Request

{  "query": "how do we ship rollout safely?",  "strategy": "hybrid_graph",  "graphDepth": 1,  "scope": {    "tenantId": "webrenew-prod",    "userId": "charles@webrenew.io",    "projectId": "github.com/webrenew/memories"  }}

Response

{  "ok": true,  "data": {    "rules": [...],    "memories": [...],    "trace": {      "strategy": "hybrid_graph",      "rolloutMode": "canary"    }  },  "error": null,  "meta": { "endpoint": "/api/sdk/v1/context/get" }}

Single Envelope

ok / data / error / meta on every SDK endpoint.

Request Tracing

Every response includes meta.requestId for support and logs.

Typed Errors

Stable error.type and error.code for deterministic client handling.

Versioned Surface

Breaking changes move to /api/sdk/v2/*, not v1 drift.

Full Endpoint Contract SDK Quickstart

Integrations

Works With Your Tools

One memory store, every coding agent. Generate native config files for each tool so your context follows you—no copy-paste, no manual sync.

The unified agent memory layer.

One memory system. Four distinct lanes.

Session

Semantic

Episodic

Procedural

Three commands. That's it.

Store, recall, generate

Built for segmented memory

Segmented Memory Architecture

Compaction-Safe Checkpoints

Session + Long-Term Recall

Tool-Native Output

Deterministic OpenClaw File Mode

Portable State

Cloud-native memory for AI apps and dev teams.

Production API that actually ships fast.

Works With Your Tools

Claude Code

Cursor

v0

GitHub Copilot

Windsurf

Gemini

Cline

Roo

OpenCode

Blackbox CLI

Codex

Amp

Kilo

Trae

Goose

OpenClaw

Antigravity

Kiro

Droid

Any MCP Client

Simple, Transparent Pricing

Free

Individual

Team Seat

Growth

Questions & Answers

Three commands. That's it.

Store, recall, generate

Production API that actually ships fast.

Simple, Transparent Pricing

Free

Individual

Team Seat

Growth

Questions & Answers