What is GhostMind?

GhostMind is an AI gateway that provides an OpenAI-compatible API for supported GhostMind capabilities.

It sits between your applications and AI provider accounts (like ChatGPT or Google Savio), providing:

• Workspace isolation — each workspace has its own connections, API keys, and usage
• Connection management — connect AI provider accounts once, share across your team
• Usage metering — track requests, tokens, and errors per workspace
• Quota enforcement — set limits on audio generation, chat, and other capabilities
• Conversation continuity — maintain context across requests with upstream conversation IDs
• Projects — organize conversations with file context and memory policies
• Multi-provider support — ChatGPT for chat, Google Savio for speech generation

How It Works

Your Application
    ↓
GhostMind API (OpenAI-compatible)
    ↓
Workspace → Connection → AI Provider Account
    ↓
AI Response → Metering → Your Application

1. You create a workspace and connect an AI provider account (a “Connection”)
2. You create an API key for that workspace
3. Your application sends requests to https://ghostmind.optdmsa.com/v1/chat/completions
4. GhostMind routes the request through your connected account
5. Usage is metered and quotas are enforced

API Surfaces

Surface	Path	Purpose
Product API	/v1/*	OpenAI-compatible API for applications
User App API	/me/v1/*	Cookie-authenticated API for the web app
Admin API	/admin/v1/*	Platform administration (private)

Developer documentation focuses on the Product API (/v1/*).

OpenAI Compatibility

GhostMind provides an OpenAI-compatible API for supported GhostMind capabilities. This means:

• The /v1/chat/completions endpoint accepts the same request format as OpenAI
• The /v1/audio/transcriptions endpoint accepts multipart audio uploads
• Streaming uses Server-Sent Events (SSE) in the same format as OpenAI

Not all OpenAI endpoints are implemented. See Models & Capabilities for the current supported set.