AI work orchestration for real servers

Projects, tasks, Git, agents, and terminals. One tight control plane.

RaccoonHQ turns scattered AI coding sessions, shell access, remote desktops, and agent conversations into managed work. Describe the outcome, bind it to a project and repository, connect an agent to your infrastructure, then watch every command, file, run, message, and handoff stay attached to the task.

1workspace for planning, execution, and audit
Liveshell, desktop, and agent chat through your connector
Githosted or external repositories per project
r5n control

workspace: tamparealestate.com website creation

task Create property website

agent TROYMIMN .25 online

bootstrap mkdir -p /var/www/tamparealestate.com

git init, remote check, first commit

instructions wrote AGENTS.md with project requirements

Ready for Codex, Claude, Gemini, or shell execution.

About

Project management wired into real infrastructure.

RaccoonHQ is a work management layer for AI-assisted engineering and operations. It keeps the plan, task queue, connector, runtime, working directory, repository, terminal transcript, remote desktop context, agent messages, and execution log in one connected place.

Instead of opening random terminals, remote desktops, and AI chats that forget the work, RaccoonHQ gives every session a project, task, repo, server, audit trail, and next step.

Launch control plane
Features

Built for real implementation work.

RaccoonHQ combines project management, live execution, and agent orchestration so users know what exists, what is blocked, what is running, and what happened.

01

Guided AI setup

Ask for the outcome in natural language. RaccoonHQ drafts the project, first task, workspace path, Git settings, AGENTS.md, and connector install flow.

02

Connector registry

Install lightweight r5n connectors on your servers, see online status, versions, update channels, host facts, and route work to the right machine.

03

Live control sessions

Open resumable PTY sessions for Codex, Claude, Gemini, OpenClaw, direct shell, or other runtimes without losing task context.

04

Remote desktop

Jump into macOS, FreeBSD, and Linux systems through browser-based remote desktop when the work needs a real GUI, not just a terminal.

05

Direct shell access

Use a clean shell path for deterministic operations: inspect files, create directories, initialize Git, restart services, and verify the host without wrapping everything in AI.

06

Text your personal agent

Use web chat or SMS to talk to your own r5n or OpenClaw agent when you are away from the dashboard. It is a private one-to-one control channel: you text your agent, and your agent replies only to you.

07

Project-aware tasks

Tasks belong to projects, carry status and ownership, and can launch control sessions with the correct connector, working directory, and thread.

08

Git and workspace bootstrap

Create directories, write instruction files, initialize or clone repositories, configure remotes, and verify Git connectivity before handoff.

09

Audit and transcripts

Terminal text, bootstrap logs, agent messages, actions, and run history are captured so work can be reviewed, resumed, debugged, or handed to another operator.

10

Consumer-safe SMS boundaries

SMS access is built for personal agent control, not campaigns: one verified consumer user per SMS number, replies only to that same verified user, no contact lists, no group blasts, no customer outreach, and no public send API to third-party recipients.

11

P2P guardrails by design

RaccoonHQ keeps personal SMS low-volume, conversational, and auditable with STOP/HELP handling, consent logs, number assignment records, message direction logs, and recipient restrictions.

12

Need more recipients?

Outbound sending from a P2P number is limited to your own verified phone number. Need additional destination numbers whitelisted? That qualifies for commercial use and requires Brand/Campaign approval and company KYC.

How it works

From intent to running agent.

RaccoonHQ keeps deterministic setup deterministic, then lets you choose the right surface for the job: AI runtime, direct shell, remote desktop, web chat, or personal SMS conversation with your own agent.

1

Describe the job

Tell the AI setup assistant what you want built, fixed, deployed, or operated.

2

Review the plan

Confirm project, repository, task, working directory, runtime, and instructions.

3

Install connector

Run one install command on the target server. RaccoonHQ waits for it to report online.

4

Execute and audit

Start AI, shell, remote desktop, web chat, or personal SMS workflows with logs attached to the task.

Pricing

Simple plans for serious work.

Start in the beta program, then grow into paid plans as your workspace, connectors, and team usage expand.

Personal

For individual builders managing their own servers.

$19/mo Get started
  • Projects and tasks
  • Single-user shell and control sessions
  • Connector install and health
  • Basic audit history

Team

For teams that need shared queues and handoff.

$49/mo Talk to us
  • Multiple users
  • Reusable Git and shell credentials
  • Hosted or external Git
  • Shared shell, desktop, and task history

Enterprise

For private infrastructure and custom deployment needs.

Custom Contact sales
  • Dedicated deployment
  • Custom retention
  • SSO and policy controls
  • Priority integration work
Contact

Need RaccoonHQ wired into your servers, desktops, Git, and AI runtime stack?

Tell us what you want to control: servers, desktops, Git, AI runtimes, or personal SMS access to your own r5n/OpenClaw agent.