Awareness Local

v1.0.0Knowledge & Memorystable

Local-first AI agent memory — one command, works offline, no account needed. Give your Claude Code, Cursor, Windsurf, OpenClaw agent persistent memory. Markdown storage, hybrid search (FTS5 + embedding), MCP protocol, Web dashboard.

ai-agentai-memoryclaude-codecursorllm
Share:
217
Stars
0
Downloads
0
Weekly
0/5

What is Awareness Local?

Awareness Local is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to local-first ai agent memory — one command, works offline, no account needed. give your claude code, cursor, windsurf, openclaw agent persistent memory. markdown storage, hybrid search (fts5 + embeddin...

Local-first AI agent memory — one command, works offline, no account needed. Give your Claude Code, Cursor, Windsurf, OpenClaw agent persistent memory. Markdown storage, hybrid search (FTS5 + embedding), MCP protocol, Web dashboard.

This server falls under the Knowledge & Memory category on MCPgee, the world's largest MCP server directory with 33,000+ servers.

Features

  • Local-first AI agent memory — one command, works offline, no

Use Cases

Create local-first agent memory
Enable offline memory storage
Build persistent agent context
edwin-hao-ai

Maintainer

LicenseMIT
Languagejavascript
Versionv1.0.0
UpdatedMay 8, 2026
Statushealthy
Maintenanceactive

Works with

ClaudeOpenAIwindowsmacoslinux

Installation

Manual Installation

npx awareness-local

Configuration

Configuration Details

Config File

claude_desktop_config.json

Performance

Response Metrics

Response Time< 200ms
ThroughputMedium

Resource Usage

Memory UsageLow
CPU UsageLow

How to Set Up and Use Awareness Local

Awareness Local is a fully offline, local-first persistent memory layer for AI coding agents. It gives Claude Code, Cursor, Windsurf, and other MCP-compatible IDEs durable memory that persists across sessions without requiring a cloud account or any external service. Memories are stored as plain Markdown files in a .awareness/ directory, indexed by SQLite FTS5 for fast keyword search and optionally augmented with embeddings for semantic recall, achieving 95.6% Recall@5 on the LongMemEval benchmark. It is ideal for long-running migrations, team handoffs, and any workflow where continuity between chat sessions matters.

Prerequisites

  • Node.js 18 or newer
  • An MCP-compatible IDE: Claude Code, Cursor, Windsurf, or OpenClaw
  • No cloud account or external API key required for offline mode
  • Optional: internet access only if enabling the cloud sync add-on
1

Run the setup command

Execute the one-command setup using npx. This installs the Awareness Local daemon, initializes the .awareness/ directory structure in your project, and registers the MCP server automatically.

npx @awareness-sdk/setup
2

Optional: enable cloud sync for cross-device use

If you want to share memory across devices or team members, re-run setup with the --cloud flag. This adds an optional sync layer while keeping the local store as the primary source.

npx @awareness-sdk/setup --cloud
3

Add the MCP server to your client configuration

Add the Awareness Local server to your MCP client config. The setup command may do this automatically, but you can also configure it manually.

{
  "mcpServers": {
    "awareness-local": {
      "command": "npx",
      "args": ["awareness-local"]
    }
  }
}
4

Initialize a session with awareness_init

At the start of each coding session, call awareness_init to load recent knowledge, open tasks, and project rules into the agent's context. This is the primary entry point for session continuity.

5

Save decisions and insights with awareness_record

During a session, use awareness_record to persist key decisions, architectural choices, and implementation notes. The daemon extracts and indexes these automatically into the knowledge store.

6

Search memory with awareness_recall or awareness_lookup

Use awareness_recall for progressive disclosure searches (summary first, then full context) or awareness_lookup for fast indexed queries over tasks, knowledge cards, and risk records.

Awareness Local Examples

Client configuration

Add Awareness Local to Claude Desktop or Claude Code config. No API keys or env vars are required for offline operation.

{
  "mcpServers": {
    "awareness-local": {
      "command": "npx",
      "args": ["awareness-local"]
    }
  }
}

Prompts to try

These prompts work once Awareness Local is running and awareness_init has been called at session start.

- "What decisions did we make about the database schema in previous sessions?"
- "Record that we decided to use PostgreSQL with row-level security for the multi-tenant feature"
- "What open tasks are left from the authentication refactor?"
- "Search my memory for anything related to the payment integration"
- "Get the agent prompt for the code-review agent so I can hand off this session"

Troubleshooting Awareness Local

awareness_init returns empty context even though previous sessions existed

Confirm the .awareness/ directory exists in your project root and that the MCP server process has read access to it. If you moved the project directory, re-run npx @awareness-sdk/setup to re-initialize the index.

Semantic search returns poor results compared to keyword search

Semantic search requires the optional embeddings index. Run the setup with embedding support enabled, or use awareness_lookup for FTS5 keyword search which works fully offline without any embedding model.

MCP server is not found after running setup

Restart your IDE or MCP client after setup completes. The setup command writes to the config file but the client must reload. If using Claude Code, run 'claude mcp list' to verify the server is registered.

Frequently Asked Questions about Awareness Local

What is Awareness Local?

Awareness Local is a Model Context Protocol (MCP) server that local-first ai agent memory — one command, works offline, no account needed. give your claude code, cursor, windsurf, openclaw agent persistent memory. markdown storage, hybrid search (fts5 + embedding), mcp protocol, web dashboard. It connects AI assistants to external tools and data sources through a standardized interface.

How do I install Awareness Local?

Follow the installation instructions on the Awareness Local GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.

Which AI clients work with Awareness Local?

Awareness Local works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.

Is Awareness Local free to use?

Yes, Awareness Local is open source and available under the MIT license. You can use it freely in both personal and commercial projects.

Browse More Knowledge & Memory MCP Servers

Explore all knowledge & memory servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.

Quick Config Preview

{ "mcpServers": { "awareness-local": { "command": "npx", "args": ["-y", "awareness-local"] } } }

Add this to your claude_desktop_config.json or .cursor/mcp.json

Read the full setup guide →

Ready to use Awareness Local?

Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.

33,000+ ServersFree & Open SourceStep-by-Step Guides