← Directory

intentic

Open Source
• Updated Aug 13, 2026, 03:11 AM

Intentic is an open-source workstation for AI agents that provides each agent with a persistent Docker sandbox on a machine you own, along with a dedicated git worktree.

Use Cases ▼
01

Run long-lived autonomous agents with persistent state

02

Provide isolated Docker sandboxes per agent for safety

03

Monitor agent activity from browser or mobile

04

Connect to agents securely via outbound-only Cloudflare tunnel

05

Maintain separate git worktrees for each agent's code

06

Self-host an agent workstation on your own infrastructure

What Problem It Solves ▼

Intentic solves the problem of ephemeral or non-persistent agent execution environments by providing each agent with a persistent Docker sandbox and git worktree, reachable via an outbound-only Cloudflare tunnel, so that agent runs can continue uninterrupted and users can monitor them from any device.

Key Features ▼
✓

Persistent Docker sandboxes ensure runs keep going

✓

Git worktree per agent provides version-controlled code

✓

Outbound-only Cloudflare tunnel enhances security

✓

Browser and mobile workspace for remote monitoring

✓

Open source and self-hosted for full control

Overview

Intentic is an open-source workstation for AI agents that provides each agent with a persistent Docker sandbox on a machine you own, along with a dedicated git worktree. It includes a browser and mobile workspace interface for monitoring and interacting with agents. Agents are reached through an outbound-only Cloudflare tunnel, which means the environment is accessible even when the client is offline, allowing long-running tasks to continue without interruption. The persistent sandboxes and git worktrees ensure that state, code, and dependencies are retained across agent runs. This design gives users full control over infrastructure while maintaining connectivity and persistence. It is particularly suited for developers and researchers who need reliable, isolated environments for autonomous agents.

Pros & Cons

Pros

+

Persistent Docker sandboxes ensure runs keep going

+

Git worktree per agent provides version-controlled code

+

Outbound-only Cloudflare tunnel enhances security

+

Browser and mobile workspace for remote monitoring

+

Open source and self-hosted for full control

Cons

−

Requires self-hosting and Docker knowledge

−

Setup complexity with Cloudflare tunnel configuration

−

Potential security risks if running untrusted agent code

−

Early-stage project with limited documentation

−

No managed cloud offering

Similar Tools

E2B Daytona Gitpod Docker Cloudflare Tunnel LangGraph

More in AI Agents