Documentation

Use EchoMemory across website, browser, mobile, IDE, and OpenClaw

Use the website for account access and memory graph entry, the browser extension for AI chat sites, the EchoChat iOS app for the native product flow, the MCP server for IDEs and desktop agents, or the OpenClaw plugin for markdown memory sync and retrieval.

See website guide Start with setup guide

Choose your setup

Most users do not need every integration. Start from the surface where you already work, then add the others only if they solve a real workflow.

Cloud web

Website

Best for sign-in, onboarding, opening the memory graph, managing API keys, and joining event or gathering graphs.

Browser

Chrome Extension

Best for ChatGPT, Claude, Gemini, Perplexity, Grok, and other supported chat sites. Capture conversations, load context, and manage memories directly in Chrome.

IDE and agents

MCP Server

Best for Cursor, Windsurf, Claude web connectors, and Claude Desktop. Gives AI tools memory search and save capabilities through MCP.

Native app

EchoChat iOS App

Best for talking with Echo directly, turning conversations into memories, browsing the community, and messaging through memory-aware profiles.

OpenClaw

OpenClaw Plugin

Best for syncing local markdown memory files into Echo cloud and enabling retrieval in Slack or OpenClaw chat flows.

Compare surfaces

Use this table to decide which guide to read first.

Surface	Best for	Runs where	Requires login	Core actions
Website	Account access, onboarding, graph entry, API key management, and event invites	EchoMemory web app	Yes	Sign in, open graph, create API keys, join gathering graphs
Chrome Extension	Saving and reusing memories from AI chat websites	Browser toolbar and content scripts	Yes for the current release	Extract memory, optimize query, load context, manage memories and sources
EchoChat iOS App	Using the native Echo product to chat, create memories, and explore people	iPhone app	Yes	Talk with Echo, generate memories, browse profiles, message another person's Echo
MCP Server	Giving IDEs and agents access to Echo memories	Local stdio process or remote MCP connector	Yes for cloud access	Search memories, save conversations, connect Cursor, Windsurf, Claude
OpenClaw Plugin	Syncing markdown memory files and retrieving them through OpenClaw	Inside OpenClaw gateway plus localhost local UI	Yes for cloud sync and search	Sync markdown, search memory, open local UI, use Slack commands

Shared prerequisites

Account1

EchoMemory account

Chrome extension usage, the EchoChat iOS app, MCP, and OpenClaw cloud sync all assume you can sign in and generate an Echo API key when needed.

Credentialec_

API key

MCP and OpenClaw depend on an Echo API key. The website guide shows where to create it, and the dedicated Get API Key page shows how to manage it safely.

ModelSurface-specific

Mode expectations

The Chrome extension is now documented as cloud-only. OpenClaw still has meaningful local-only behavior in its localhost UI, so do not merge those two ideas.

Start with Choose Your Setup if you are not sure which integration matches your workflow.

Recommended paths

I need the main website flow first

Read Website. That guide covers sign-in, onboarding, opening the memory graph, API key entry points, and event or gathering invites.

I want EchoMemory in browser chats

Read Chrome Extension first. Add the MCP guide later only if you want the same memories available inside your IDE or Claude.

I want the native Echo experience on my phone

Read EchoChat iOS App. That page now includes the walkthrough videos and the screenshot-based product tour that used to live on the separate demo page.

I want my coding agent to remember past work

Read MCP Server. That is the cleanest path for Cursor, Windsurf, Claude web connectors, and Claude Desktop.

I want to sync markdown notes from OpenClaw

Read OpenClaw Plugin. That guide covers local UI, sync behavior, Slack commands, and config.

I need an API key for MCP or OpenClaw

Read Get API Key before touching either config guide.

I am already stuck

Jump straight to Troubleshooting and then use FAQ for quick decision questions.