Looped AgentsAlpha

Paste the file. Ship the agent.

The hosted home for Looped agent-framework agents. Paste the same agent.yaml you run locally, store its secrets, and deploy — each agent gets its own isolated microVM, wakes on its triggers, and sleeps when idle.

Free tier · No infrastructure to manage · Built on looped-af

agents.looped.sh
standup-botcron · Mon–Fri 9:00healthy
issue-triagegithub · webhooksleeping
support-botdiscord · livedeploying
overview · config · secrets · runs · audit · logs · terminal

Deployment

The same file you run locally.

No hosted-only config, no rewrite. The agent.yaml you built with the open-source framework deploys unchanged — validated first, then provisioned onto its own isolated microVM with an encrypted volume and the hardened base image.

  • One isolated microVM and encrypted volume per agent
  • Deploys are durable workflows — crashes resume, never half-deploy
  • Edit the config and it flags exactly what needs a redeploy
Deploying standup-botdeploying
  • Validate agent.yaml
  • Resolve secrets
  • Provision isolated microVM
  • Health check

Durable workflow — a crashed deploy resumes where it stopped.

Discord
Slack
Telegram
Email
Webhook
GitHub
Cron
TTY

Any trigger wakes a sleeping agent. Cron schedules wake it on time.

Scale to zero

Asleep until something happens.

Hosted agents spend most of their lives waiting. So by default they sleep — and a Slack mention, a webhook, an email, or a cron tick wakes them in seconds. You pay attention to what they did, not to keeping them alive. Checking the dashboard never wakes a sleeping agent.

  • Sleeps after idle, wakes on any trigger
  • Always-on mode for live Discord voice and long sessions
  • Free-tier models routed through the platform, metered per agent

Trust

Locked down and written down.

Secrets are encrypted and write-only — the dashboard shows names, never values. Everything the agent does lands in its runs, audit trail, and logs, and when you need to get hands-on there's a live terminal into the running machine.

  • Write-only secrets, encrypted at rest
  • Runs, audit trail, and logs for every agent
  • A live terminal, one click away
SecretsEncrypted · AES-256-GCM
DISCORD_BOT_TOKENdiscord trigger••••••••write-only
GITHUB_TOKENreferenced in agent.yaml••••••••write-only
OPENAI_API_KEYbring your own key••••••••write-only

There is no API to read a secret back out. Not even yours.

Leave the hosting to us

Seconds to deploy. Not hours.

You built the agent — that was the interesting part. Don't spend the rest of the day on Docker, compose files, and a VPS to babysit. Paste the file and we provision it, keep it patched, wake it, sleep it, and watch it for you.

Hosting it yourself

  • Provision a box
  • Write the compose file
  • Wire up TLS and ports
  • Ship the logs somewhere
  • Keep it patched and alive

An afternoon, if it goes well.

Looped Agents

  • 1Paste your agent.yaml
  • 2Add its secrets
  • 3Deploy

Seconds later: healthy, and on the clock.

Prefer to run things yourself? Everything underneath is open source — the agent framework and Looped Meet are self-hostable, Apache 2.0.

Free tier · Bring your own key for more models

Deploy your first agent today.

Talk to a human

Got something Looped can't handle yet?