Use Cases

Agentic work where your infrastructure already is.

Use PawFlow when reasoning, real-machine access, persistent context, and dependable automation need to work together.

AI coding

Code on your own machines

Give an agent controlled access to the real repo, toolchain, terminal, browser, and local services without moving the project to a vendor sandbox.

Operations

Route work through Relays on a workstation, server, GPU host, or remote environment from one shared runtime.

Self-hosted workspace

Use the same agents, memory, skills, tools, permissions, and conversations across Web, CLI, VS Code, Android, and API clients.

Automation

Let an agent discover a process, then encode its repeatable execution in a visual Flow with retries and checkpoints.

UI automation

Use screen, browser, desktop, and noVNC capabilities through a Relay when the work cannot be completed through an API alone.

Media

Connect vision, image, video, audio, voice, 3D, ComfyUI, and GPU services to agent and Flow work.

Private AI

Self-host the control plane, scope Relay tools, gate sensitive actions, and choose where models and secrets are resolved.

External AI clients

Let ChatGPT or another compatible client inspect current project data through a read-only route instead of an uploaded archive.

Start with one real task

The shortest path to understanding PawFlow is to use it on a workspace you already know.