AgentHost gives developers who use multiple AI agents one persistent Control Room—with shared memory, coordinated work, governed handoffs, and a durable record in infrastructure they control.
the governance layer — patent pending (US provisional 64/122,926) · five engines on your box: Claude Code · Codex · Gemini CLI · Kimi K3 · Hermes — plus OpenClaw for Telegram & DiscordFive agents review a pull request on one box. The red-team agent finds a concurrency hazard and fails it. A hotfix goes in, the council re-runs, and it merges — under the strictness level the operator set.
The council in this clip is scripted. The real one streams live from your box.
One canonical board, a developer workspace, and five governed engines on a box you own — reachable from any device over your private mesh. Underneath it, the delivery pipeline every change walks before it touches the live box.
One control plane. Many intelligent entities. Every change earns its way to the live box.
Your agents keep their memory, tools, and work state across sessions — with budgets, approvals, and audit trails you control.
Claude Code and Codex can execute and cross-review approved autonomous tasks. Hermes Agent and Gemini CLI join chat workflows. OpenClaw runs separately as the Telegram and Discord gateway. Each arrives with its identity intact — skills, memory, MCP servers, plugins — on the same box.
Spend caps per work chain. Human approval gates on deploy, delete, send, and credentials. Cross-engine review before work is marked done. Every action audited. Deploying, spending, deleting, email or messages, and ambiguous work wait for a human. Social publishing can run unattended only through the narrow Postiz allowlist.
The first live autonomy day surfaced four real bugs on the founder's production box — and the governance system caught every one.
Tasks stopped at the gates, budgets capped runs, and cross-engine review rejected bad results before they reached production.
The CLI is public on npm. The production log is the proof until real customers exist.
Anyone can start a cloud container. The hard part is turning a session-bound tool into a persistent, governed engineering system. That's what AgentHost does.
Watch one task move. The board keeps its owner, state, reviewer, and next decision visible. When the work needs you, it lands in Awaiting Your Review and the phone alert opens the exact card.
A 7am briefing reads your repos and memory and hands you a digest before you're up.
The box sends them itself — no vendor backend required.
Roles, handoff format, shared-board behavior, budgets, and blocked-work reporting load for Claude, Hermes, and Codex — persistent across sessions.
Eligible Claude and Codex tasks run in restricted environments. Risky or ambiguous work parks for you, and every chain has hard ceilings.
Upload files to the box, download finished artifacts, or send text back to your desktop through the optional bridge.
Skills, memory, MCP servers, plugins — all migrate. Keep Claude Code, Codex, Hermes Agent, Gemini CLI, OpenClaw, MCP, memory, and messaging on one box, with the Control Room above the work. AgentHost adds persistent infrastructure; supported tasks gain budgets, review, and a durable record.
Real product marks identify supported tools and channels. They do not imply endorsement, sponsorship, or partnership.
Superpowers, AgentMemory, and CodeGraph for planning, testing, review, memory, and local code context.
includedClaude Ads, Claude SEO, Claude Blog, and Marketing Skills for research through distribution.
includedNewsjacking and Pitch Deck Mastery arrive as standalone skills you can invoke immediately.
includedHiggsfield is present for image and video generation when you add its service key.
key neededSome Loops need the matching email, calendar, repository, or vault connection before they can run.
AgentHost brings your existing skills, memory, MCP servers, plugins, and project state first. Your agent arrives as itself. Opt in to carry your MCP logins too. The starter stack fills gaps without clobbering your choices.
Your agent arrives with its identity intact. Credential files never migrate. Auth values you explicitly supply go straight from your laptop to your cloud provider's encrypted store. No AgentHost backend receives them. Your cloud account, your repos, your model keys, your spend controls. Your agent survives lid-close, sleep, and airport wifi.
Your cloud account, your credentials, your repos, your model keys, your spend controls. Spend caps per work chain. Human approval gates on deploy, delete, send, and credentials. Cross-engine review before work is marked done. Every login, run, and job audited. The value never appears in chat history or returns to the interface. Autonomous runs receive no box secrets except the narrowly allowlisted POSTIZ_API_KEY used for supported social-publishing tasks.
Your cloud account, credentials, repos, model keys, task history, memory, budgets, and review rules live in your environment. Models and subscriptions can change without forcing you to rebuild the operation. Bring-your-own-cloud is the default.
Both paths put the runtime in your cloud account. Your provider and model costs stay separate.
AgentHost fee. Your cloud and model providers bill you separately; configuration determines the cloud cost.
$49 AgentHost fee, held for life while continuously subscribed. $99/mo after the Founding 50. Your cloud and model providers remain separate — about the cost of one Claude Max seat, and it puts everything you already pay for to work around the clock, governed.
agenthost destroy. The app, volume, and secrets are removed from your cloud account.