Epicenter is a local‑first personal data platform that stores all notes, transcripts, and other content as plain‑text files and SQLite databases in a single folder you own. It lets you edit and search your data with existing tools like Obsidian or VS Code and integrates with any LLM via an API key, enabling context‑aware AI queries without copy‑pasting or cloud lock‑in. The open‑source system includes a local OpenCode instance and optional secure tunneling for remote access, ensuring privacy and extensibility.
Funding
Funding not disclosed
Founders
Product
Problem
Users of multiple productivity and AI tools often store data in separate, proprietary services, leading to fragmented workflows, data lock‑in, and difficulty sharing context across applications.
Solution
Epicenter provides a local‑first personal data platform built on plain‑text files and SQLite databases that serves as a single source of truth for all user content. By organizing everything in one folder, the system lets transcripts, notes, and other artifacts be edited with existing tools such as Obsidian, VS Code, or command‑line utilities without any copy‑pasting or export steps. The platform integrates with any LLM via an API key, allowing AI assistants to query the unified workspace and generate context‑aware responses. All data remains on the user’s machine unless they explicitly enable syncing, ensuring privacy and eliminating cloud lock‑in. Open‑source components, including a local OpenCode instance and secure tunneling, enable developers to extend the ecosystem while preserving performance through Rust‑based architecture.
Target Audience
Primary users are knowledge workers, researchers, and developers who rely on multiple note‑taking, transcription, and AI tools and seek a unified, privacy‑preserving workspace.
Features
- Single‑folder architecture storing all content as plain‑text files and SQLite databases that users own
- Local‑first design with optional secure tunneling for remote access
- Compatibility with existing tools (Obsidian, VS Code, grep/ripgrep, Git) for seamless editing, searching, and versioning
- Model‑agnostic AI integration supporting OpenAI, Anthropic, or self‑hosted LLMs via user‑provided API keys
- OpenCode local instance for running AI queries directly against the workspace
- High‑performance core written in Rust and Svelte 5, leveraging CRDTs for real‑time collaboration
- Fully open‑source codebase allowing auditing, forking, and custom extensions