HeronMulti-agent orchestration
Each agent is a staff member with one clear role (PM, BA, developer, QA, security…). Heron builds the tasks, dependencies and handoffs between agents, with verification evidence for every step.
Free sign-up is open
Keep devices and AI agents ready — open control only when it is really needed
Heron brings project-based AI agent orchestration and device administration into one system: tasks, dependencies, evidence-backed handoffs, 24/7 device presence and small, audited control capabilities.
Ask Heron AI
What Heron does
Secure remote device management
Devices pair with a one-time code, send a heartbeat every 30 seconds and show online/stale status per organization. No inbound ports on user machines: the agent only connects outward.
Local-first
Private keys and execution rights stay on your device. The cloud is only a control plane for sign-in, organizations, devices and audit.
Small, approved, audited capabilities
Instead of an all-powerful remote shell, Heron grants small policy-bound capabilities, asks for approval when needed and audits every action.
Real-time collaboration
Chat rooms and group meetings with AI recaps (summary, decisions, action items) are in development, with data isolated per organization.
AI cost control
Usage is tracked per account and model with quotas; Heron never switches you to another model without your consent.
How Heron works
- Create an organization and invite members as Owner, Admin, Member or Viewer.
- Pair devices with a one-time code; the Heron Device Agent runs directly on the machine and only connects outward.
- Describe a project; Heron drafts the backlog, picks agents by expertise and runs a dependency-aware workflow.
- Follow progress, approve important decisions and review the evidence before accepting results.
Who it is for
- Developers and AI power users who monitor their own machines and coordinate coding agents.
- Software teams of 3–20 people who need RBAC, audit, approvals and shared tasks.
- MSPs, labs and mid-sized companies managing many devices, policies and audit logs.
Security
- Data is isolated per organization with PostgreSQL row-level security.
- Passwords and pairing codes are stored as salted hashes; device tokens are shown only once.
- The admin area sits behind Cloudflare Access on enrolled devices.
- Per-request Content Security Policy, Turnstile bot protection and rate limiting.
Guides to AI agents
What is an AI agent? How it differs from a chatbot and what it is for
An AI agent is software that uses a large language model to plan, call tools and finish a multi-step goal. This guide explains the parts, examples, risks and how to get started.
Multi-agent orchestration: roles, dependencies and evidence-backed handoffs
How to make several AI agents work like a software team: split PM, BA, developer, QA and security roles; manage dependencies; hand off with evidence; and keep humans approving the important decisions.
Claude Code vs Codex: which coding agent to choose, and using both
A practical comparison of two terminal coding agents — Anthropic's Claude Code and OpenAI's Codex: sign-in, how they work on a codebase, when to use which, and how to orchestrate both in one project.
AI agent security: least privilege, approvals and audit when agents run on your machine
A security checklist for letting AI agents act on real machines and systems: least privilege, outbound-only connections, human approval, audit trails, secret handling and prompt-injection defence.
AI agent cost control: quotas, model choice and usage tracking
Why AI agent costs overrun and how to control them: quotas per account and model, choosing models by kind of work, trimming context, never switching models without consent and measuring cost per accepted result.
Remote device management for AI agents: pairing, heartbeats and outbound-only connections
How to manage the machines your AI agents run on safely: one-time pairing codes, heartbeats to know which machine is online, outbound-only connections and permissions per organization and role.
Frequently asked questions
What is Heron?
Heron is a platform for project-based AI agent orchestration and secure remote device management, built for developers, software teams and small businesses.
Does Heron need inbound ports on my machine?
No. The Heron Device Agent only opens outbound connections to Heron; your machine accepts no inbound connections.
Can Heron run arbitrary commands on my device?
No. Heron uses small policy-bound capabilities with approval and audit instead of a full remote shell.
How can I try Heron?
Create a free account at guday.space/register. You join an organization through an owner's invite link, or request a new organization for a Heron admin to approve.
How are Heron and Guday Space related?
Heron is built by Guday Space (guday.space) and is also known as Heron by Guday Space.