AgentFS provides a SQLite-based agent filesystem with a Rust core, an SDK in multiple languages, and a CLI for mounting and managing agent state. It stores all agent activity in a single SQLite database file and offers filesystem, key-value, and toolcall interfaces.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
AgentFS is a filesystem explicitly designed for AI agents. It provides three interfaces for agent state management: a POSIX-like filesystem, a key-value store, and a toolcall audit trail. The entire agent runtime—files, state, history—is stored in a single SQLite database file, implemented with Turso.
How it works
AgentFS stores all agent data in a SQLite database and offers:
- Filesystem operations (read/write/list) via an SDK (TypeScript, Python, Rust) and a CLI for mounting (FUSE on Linux, NFS on macOS).
- A key-value store API for agent state and context.
- A toolcall audit trail for debugging and analysis. The database can be inspected, snapshot, and moved between machines.
Getting started
Getting started sections in README include:
- Getting started with the CLI:
- Install the AgentFS CLI:
curl -fsSL https://agentfs.ai/install | bash - Initialize an agent filesystem:
$ agentfs init my-agent Created agent filesystem: .agentfs/my-agent.db Agent ID: my-agent
- Install the AgentFS CLI:
- Inspecting the filesystem:
$ agentfs fs my-agent ls Using agent: my-agent f hello.txt $ agentfs fs my-agent cat hello.txt hello from agent - Mounting the filesystem:
$ agentfs mount my-agent ./mnt $ echo "hello" > ./mnt/hello.txt $ cat ./mnt/hello.txt hello - Running programs in a sandbox:
$ agentfs run /bin/bash Welcome to AgentFS! $ echo "hello from agent" > /agent/hello.txt $ cat /agent/hello.txt hello from agent $ exit - SDK usage:
npm install agentfs-sdkimport { AgentFS } from 'agentfs-sdk'; const agent = await AgentFS.open({ id: 'my-agent' }); const ephemeralAgent = await AgentFS.open(); await agent.kv.set('user:preferences', { theme: 'dark' }); const prefs = await agent.kv.get('user:preferences'); await agent.fs.writeFile('/output/report.pdf', pdfBuffer); const files = await agent.fs.readdir('/output'); await agent.tools.record('web_search', Date.now() / 1000, Date.now() / 1000 + 1.5, { query: 'AI' }, { results: [...] });
Recent releases
- v0.6.4 (2026-03-25): Fixed TypeScript SDK dev dependency to fix CI build for serverless. Installation snippet:
cur - v0.6.3 0.6.3 - 2026-03-25 (2026-03-25): ## Release Notes ### Added - TypeScript SDK: Serverless adapter for `@tursodatabase/serverless`. - v0.6.2 (2026-02-21): Fixed native-tls update to fix nightly build.
- v0.6.0 (2026-02-05): AgentFS is now beta and added migrate/exec commands and -c option to init.
Traction
Stars: 3320 Forks: 184 Open issues: 72
License
MIT






