ai-filesystem 0.1.4

Virtual filesystem service for AI agents with LangChain integration

AI Filesystem

A virtual filesystem API for AI agents. Users can easily provide their agents with siloed filesystems to use as long term memory, scratchpads, or general purpose storage.

How It Works

1. API Server: FastAPI backend that stores files in PostgreSQL
2. User Isolation: Each user can only see/modify their own files (enforced by database)
3. Multiple Filesystems: A user can have multiple filesystems and provide their agents with access to different, or the same filesystems.

[Beta] Quick Start

Creating an account on the Filesystem

1. Navigate here and sign up for an account on the Filesystem.
2. Create an API key and save it someplace secure!

Giving your agent access to the Filesystem

1. Set your API key as an environment variable for your agent: AGENT_FS_API_KEY=<api_key>
2. Specify the URL for the filesystem. If you're using our hosted solution, it is AGENT_FS_URL=agent-file-system-production.up.railway.app
3. Instantiate the filesystem client, and give your agent access to the tools

from ai_filesystem import FilesystemClient

client = FilesystemClient(
  filesystem="nicks-agent-filesystem"
)
filesystem_tools = client.create_tools()  # list files, read file, create new file, and edit file
agent.bind_tools(filesystem_tools)

3. In your agent's system prompt, make sure to specify how you want the agent to use the filesystem. Common use cases include as long-term memory, to store learnings and mistakes, or to save work products.