ChatGPT Integration

Use OpenAI tools with ACE - Codex CLI terminal tool or ChatGPT Custom GPTs for persistent memory

What You'll Build

OpenAI integration with ACE via Codex CLI or Custom GPTs

Two ways to use OpenAI tools with ACE:

OpenAI Codex CLI: Terminal tool that executes bash commands locally to call ACE's REST API

ChatGPT Custom GPTs: Web-based ChatGPT with Actions calling ACE's API (requires public URL or tunnel)

Track All 9 Entity Types: Issues, Decisions, Work Logs, Architecture, Best Practices, Patterns, Tags, Relationships, Memories

Knowledge Graph: Create and traverse relationships between entities

Prerequisites

What you'll need

ChatGPT Plus or Pro

Required for Custom GPTs and Codex coding agent access

ACE Installed and Running

Install ACE and start the server locally

Namespace Created

Create a namespace in the ACE dashboard

API Keys (Optional)

Configure OpenAI key for semantic search. Learn about BYOK →

Quick Start

Get ACE server running locally

1. Install and Start ACE:

# Install ACE (requires Node.js 18+)
npm install -g @ace3-memory/ace

# Login and configure
ace login  # Opens browser for OAuth
ace init   # Configure database connection
ace start  # Runs on localhost:7777

2. Test the Server:

curl http://localhost:7777/health

Should return: {"status": "healthy", "service": "ACE Universal Server"}

3. Get Your API Key:

ace api-key create --name "ChatGPT Integration"

Save this API key - you'll need it for Custom GPT Actions or when configuring Codex agent.

That's it! Your ACE server is running. Continue below for usage examples.

REST API Examples

Common operations for OpenAI Codex CLI

Note: Replace {YOUR_NAMESPACE} with your namespace from the dashboard. These examples use localhost where ACE runs with ace start.

Track an Issue:

curl -X POST http://localhost:7777/api/v1/{YOUR_NAMESPACE}/issues \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Fix authentication bug",
    "description": "Users can't log in with OAuth",
    "status": "open",
    "priority": "high",
    "tags": ["security", "auth"]
  }'

Log a Decision:

curl -X POST http://localhost:7777/api/v1/{YOUR_NAMESPACE}/decisions \
  -H "Content-Type: application/json" \
  -d '{
    "decision": "Use PostgreSQL for production",
    "rationale": "Better performance and ecosystem support",
    "alternatives": ["MySQL", "MongoDB"],
    "tags": ["architecture", "database"]
  }'

List and Filter Issues:

curl "http://localhost:7777/api/v1/{YOUR_NAMESPACE}/issues?status=open&limit=10" \
  -H "Content-Type: application/json"

Create Work Log:

curl -X POST http://localhost:7777/api/v1/{YOUR_NAMESPACE}/work-logs \
  -H "Content-Type: application/json" \
  -d '{
    "date": "2025-11-18",
    "summary": "Implemented user authentication",
    "tasks_completed": ["Added JWT tokens", "Created login page"],
    "hours": 4,
    "tags": ["authentication"]
  }'

💡 Optional: Environment Variables for Convenience

If you prefer, you can set an environment variable to avoid repeating your namespace:

export ACE_NAMESPACE="your-namespace"

Then use $ACE_NAMESPACE in commands:

curl http://localhost:7777/api/v1/$ACE_NAMESPACE/issues

Knowledge Graph

Link entities together

Create a Relationship:

curl -X POST http://localhost:7777/api/v1/{YOUR_NAMESPACE}/relationships \
  -H "Content-Type: application/json" \
  -d '{
    "source_type": "decision",
    "source_id": 1,
    "target_type": "architecture",
    "target_id": 2,
    "relationship_type": "implements",
    "metadata": "{\"phase\": \"initial-setup\"}"
  }'

Traverse the Graph:

curl "http://localhost:7777/api/v1/{YOUR_NAMESPACE}/relationships/traverse/decision/1"

How ChatGPT Uses ACE

Example workflow with Codex agent

When using ChatGPT's Codex coding agent, you can ask it to track work, log decisions, or query project information. Codex will execute API commands against your ACE server.

YOU SAY:

"Log the decision to use PostgreSQL for our database"

↓

CODEX AGENT EXECUTES:

curl -X POST http://localhost:7777/api/v1/{YOUR_NAMESPACE}/decisions \
  -H "Content-Type: application/json" \
  -d '{
    "decision": "Use PostgreSQL for production database",
    "rationale": "Better performance, ACID compliance, ecosystem",
    "alternatives": ["MySQL", "MongoDB"]
  }'

↓

CODEX RESPONDS:

"I've logged your decision to use PostgreSQL. Decision ID: 5"

🌐 Alternative: Custom GPT (Web-Only)

For users who want a ChatGPT web interface

If you prefer using ChatGPT in the browser instead of CLI tools, you can create a Custom GPT with ACE Actions. This requires ChatGPT Plus ($20/month) and works only in the ChatGPT web interface.

Quick Setup:

Go to ChatGPT GPT Editor
Create a new GPT with a name like "ACE Memory Assistant"
In Actions, import ACE's OpenAPI spec from Dashboard → API Reference
Set API Key authentication with your ACE API key
Save and test by asking it to create an issue or decision

Note: The Codex coding agent can execute API commands without Custom GPTs. Custom GPTs are useful if you want a persistent chat assistant with built-in ACE memory Actions.