Versions:

0.4.7
0.4.6
0.4.5
0.4.4
0.4.3
0.4.2
0.4.1
0.4.0
0.3.8
0.3.7
0.3.6
0.3.5
0.3.4
0.3.3
0.3.1
0.3.0
0.2.8
0.2.7
0.2.6
0.2.5

MindFS, published by a9gent and currently at version 0.4.7 with 20 recorded versions, is a self-hosted tool for accessing personal AI agents and workstation data from anywhere. It falls into the AI assistant access and remote file management category, unifying agent sessions and project files behind a single web interface. Its agent session features support multiple agents—including Claude Code, OpenAI Codex, Gemini CLI, Cursor, Copilot, Cline, Augment, Kimi, Kiro, Qwen, Qoder, Pi, OpenCode, and OpenClaw—with automatic detection of installed agents, token-by-token streaming output rendered as structured cards for tool calls, reasoning, and permission requests, mid-session switching between agents or models over shared context, session search, bidirectional import and synchronization with existing CLI sessions, persistent binding recovery after restarts, rich media input, and multi-device synchronization. File access capabilities include hosting multiple projects, storing all data in a .mindfs/ subdirectory for easy migration and backup, a full file tree with renderers for Markdown, images, and code, and git status and worktree support. Interaction features cover slash commands, @ file references, # quick prompts, and bidirectional navigation between files and sessions, with PWA installation for desktop and Android and a mobile-optimized interface. Access modes include local LAN browser access without accounts, a Relay remote mode requiring no open firewall ports, private channel access such as Tailscale, and end-to-end encryption for sessions and files. A plugin system allows custom file views that agents themselves can generate. It runs as a single statically compiled binary under 10 MB with zero dependencies, supporting macOS (Intel and Apple Silicon), Linux (x86-64, ARM64, ARMv7), and Windows.

Tags: