Universal Memory Infrastructure for AI

Name: ACE3
Author: ACE3

9 entity types. Knowledge graph. 20 MCP tools. Works with ChatGPT, Claude, Gemini, Cursor, and any AI.

No credit card required • 100 free memories • 50 API calls/day

20 MCP tools

60+ REST endpoints

Sub-100ms response

What Makes ACE Different

Not just storage - a complete memory infrastructure with structured entities, relationships, and intelligent retrieval

9 Entity Types

Not just notes - structured Issues, Decisions, Architecture, Patterns, Best Practices, Work Logs, and more. Each with full context.

Knowledge Graph

12 relationship types. Traverse connections. Find how a decision led to a bug and discover its solution automatically.

MCP Server

20 native tools for Claude Code, Gemini CLI, Codex CLI, Cursor, Windsurf. Your AI calls ace_remember, ace_track_issue directly.

AI Builder

Generate complete project plans from requirements. Creates decisions, architecture, patterns, issues - all linked together.

Semantic Search

AI finds relevant context instantly. Not keyword matching - true semantic understanding with OpenAI embeddings + pgvector.

Your Infrastructure

Data stays on your PostgreSQL. Neon free tier, Supabase, AWS RDS, or local. Zero telemetry. Complete sovereignty.

Native AI Integration via MCP

The Model Context Protocol gives your AI direct access to ACE memory. No HTTP requests needed - your AI calls tools natively.

20 specialized tools - memory, issues, decisions, graph traversal, AI planning
9 read-only resources - direct data access
6 workflow prompts - code review, session management, bug tracking

// claude_desktop_config.json

{
  "mcpServers": {
    "ace": {
      "command": "ace",
      "args": ["mcp", "--namespace", "my-project"]
    }
  }
}

Works with Claude Code, Gemini CLI, Codex CLI, Cursor, Windsurf, and any MCP client

More Than Memory - It's a Knowledge Graph

Connect entities with semantic relationships. Your AI traverses the graph to find context you didn't know you needed.

12 Relationship Types

implementsreferencessupersedesrelates_tocaused_byresolvesdocumented_independs_onblocksduplicatesvalidatesreproduces

Intelligent Queries

"What issues are blocked by this decision?"
"Find all bugs caused by this component"
"Show patterns that implement this decision"
"Traverse from bug → root cause → fix → prevention"

Example input:

"Build a user authentication system with OAuth2, JWT tokens, and role-based access control for a Next.js app"

Generated output:

3 architectural decisions
5 architecture components
4 best practices
2 design patterns
All linked with relationships

Describe It. ACE Builds It.

Generate complete project plans from natural language. ACE creates structured entities and links them automatically.

Natural language to structured entities
Automatic relationship linking
Review before import
Works with GPT-4, Claude, or Gemini

Get Started in 3 Steps

From zero to AI memory in under 2 minutes

Install

npm install -g @ace3-memory/ace

Login

ace login

Start

ace start

Your AI now has persistent memory. Every decision, pattern, and context is preserved and searchable.

Works With Any AI

All major AI tools support MCP. ACE also provides a REST API for universal access.

🧠

Claude CodeMCP + REST

🤖

Codex CLIMCP + REST

✨

Gemini CLIMCP + REST

📝

CursorMCP + REST

🌊

WindsurfMCP + REST

💬

ChatGPTMCP + REST

🚀

CopilotREST API

🔧

Any AIREST API

MCP Server (20 tools) for native integration • REST API (60+ endpoints) for universal access