Turnstone

Product Introduction

1. Definition: Turnstone is a local-first, AI-native workspace and agent orchestration platform. Technically, it is a desktop application that functions as a unified knowledge graph, integrating data from both cloud applications and local file systems to create a persistent, personalized AI context.
2. Core Value Proposition: Turnstone exists to eliminate AI context amnesia. Its primary value is creating a persistent "second brain" that learns from a user's entire digital footprint, enabling specialized AI agents to operate with deep, continuous personal context without requiring repeated manual prompting or file uploads.

Main Features

1. Unified Knowledge Graph & Local-First Architecture: Turnstone creates a private, searchable index of all connected data. It connects to SaaS apps (Gmail, Google Calendar, Slack, Google Drive) and local folders on your machine. All data processing and indexing occur locally on the user's device. The AI agents query this local knowledge graph, meaning sensitive data never leaves the user's control unless explicitly sent to a cloud AI model for processing.
2. Context-Aware AI Agent Orchestration: Users can create and deploy specialized agents (e.g., Chief of Staff, Research Analyst, GTM Strategist). Crucially, all agents share access to the same central Turnstone knowledge graph. This means an agent tasked with drafting an email has implicit access to relevant calendar events, past communications, and related documents without manual context provisioning.
3. Bring-Your-Own-Model (BYOM) Engine: Turnstone is model-agnostic. It provides the orchestration layer and context, but the AI reasoning can be powered by any API-compatible large language model (LLM) the user already pays for or prefers, such as OpenAI's GPT-4, Anthropic's Claude, or local models via Ollama. This decouples the agent's intelligence from the platform's infrastructure.

Problems Solved

1. Pain Point: The high cognitive overhead and inefficiency of manually providing context to every new AI chat session. Traditional AI tools treat each interaction as a blank slate, forcing users to re-upload files and re-explain background, preferences, and ongoing projects repeatedly.
2. Target Audience: Knowledge workers, executives, entrepreneurs, and researchers who manage complex projects across multiple applications. Specific personas include startup founders, product managers, consultants, and content creators who spend significant time synthesizing information from emails, documents, and meetings.
3. Use Cases: Preparing for a board meeting by having an agent synthesize the last quarter's emails, Slack discussions, and strategy documents into a briefing. Conducting competitive research where an agent analyzes saved PDFs, recent news, and internal memos. Drafting personalized client communications that reference past interactions and project details automatically.

Unique Advantages

1. Differentiation: Unlike cloud-only AI assistants (like ChatGPT or Gemini) which lack persistent personal context, or fragmented productivity suites, Turnstone offers a unified, persistent memory layer. Unlike enterprise knowledge bases, it is personal, automatic, and agent-ready.
2. Key Innovation: The combination of a local-first knowledge graph with a multi-agent, shared-context architecture. The technical innovation is treating the user's disparate data as a single, queryable source of truth for multiple AI agents, all while maintaining privacy and model flexibility through its BYOM and local processing design.

Frequently Asked Questions (FAQ)

1. Is my data private with Turnstone? Yes, Turnstone employs a local-first architecture. Your data from apps and folders is indexed and stored locally on your machine. Your data is only sent to external servers when you explicitly query an AI model (like OpenAI), and even then, only the necessary context for that query is sent.
2. What AI models does Turnstone work with? Turnstone is compatible with any LLM that offers an API, including OpenAI GPT, Anthropic Claude, and open-source models running locally via tools like Ollama. You configure your own API keys, so you use and pay for the models you choose.
3. How does Turnstone connect to my apps like Gmail and Slack? Turnstone uses secure OAuth connections to read data from your connected cloud applications. This is a standard, read-only connection used to sync your data to the local knowledge graph. You can revoke access at any time.
4. Can I use Turnstone for team collaboration? Currently, Turnstone is focused on the individual knowledge worker as a personal AI workspace. Its core architecture is designed around a single user's unified context, not real-time multi-user collaboration.
5. What is the difference between an AI agent and a chatbot in Turnstone? A Turnstone agent is a persistent, specialized instance (e.g., a "Research Agent") with ongoing access to your entire knowledge graph. A chatbot is typically a single-session, context-less interaction. Agents in Turnstone have memory, purpose, and deep context built-in.