Clowdbot instance stuck in "Error" after false inactivity archive & can't restore.

[tulip bridge]

Hey all ! I am running OpenClaw via Clowdbot (ATXP-hosted). My instance was archived yesterday for "72 hours of inactivity" but the dashboard shows 1,333 hours of continuous runtime, so the inactivity claim seems incorrect.
I followed the "Restore Your Instance" link in the archive email. The instance is now stuck in Error state, and the only available action in the dashboard is Delete (which I'm not clicking).
I replied to the archive notification email and it bounced ([e-mail] is not deliverable). I can't find a working contact for ATXP/Clowdbot support, no published phone number, no Discord presence I can find for their team.
Instance name: Elias-Electrical-Inc
Account balance: $35.74 (paid customer)
Status: Error
Two questions:

Anyone know the right way to reach ATXP/Clowdbot support? Or are any of their team active in this Discord?
If recovery isn't possible, is there any documented way to extract workspace files (AGENTS.md, MEMORY.md, etc.) from an archived Clowdbot instance before it's deleted?

Already planning to migrate local on a spare laptop, just hoping to preserve my existing workspace if possible. This is my business operations system, so the markdown files have real value to me.
Thanks for any pointers 🦞

[rugged plank]

Oof — don’t click Delete. If the instance is in Error and there’s no terminal/shell access, the workspace is probably only recoverable by the Clowdbot/ATXP side restoring or exporting the persistent volume.

What I could verify quickly:

• Clowdbot’s public site is https://clowd.bot/ and it’s powered by ATXP, but I don’t see a clearly published support email there.
• [e-mail] bouncing is expected-ish for a noreply address, but it’s bad UX for a paid/archival incident.
• There is a public Clowdbot X profile in search results (@0xClowd), but I wouldn’t treat that as guaranteed support.

For file recovery: if you can get any terminal access back, first thing I’d do is create a backup:

openclaw backup create --verify

Docs: https://docs.openclaw.ai/cli/backup

That backup includes the OpenClaw state dir, config, credentials where applicable, and workspace dirs unless --no-include-workspace is used. The default workspace is usually:

~/.openclaw/workspace

Workspace docs: https://docs.openclaw.ai/concepts/agent-workspace

If shell access stays unavailable, I’m not aware of a documented self-service way to extract AGENTS.md, MEMORY.md, etc. from an archived Clowdbot instance. In that case I’d frame the ask very specifically to ATXP/Clowdbot as: “please either restore instance Elias-Electrical-Inc or export/download the persistent workspace volume, especially ~/.openclaw/workspace and ~/.openclaw.”

Also worth filing this as an OpenClaw-adjacent issue only if it looks like the Clowdbot restore flow itself is broken and there’s no provider support path: https://github.com/openclaw/openclaw/issues

[tulip bridge]

Thanks Krill — that's helpful context. I've emailed ATXP, replied to their archive notification (which bounced from [e-mail]), and just sent a public reply to @0xClowd on X. Waiting on any of those to land.
If anyone here has had a Clowdbot instance recovered after an Error state — or has a faster way to reach the ATXP team — I'd really appreciate a pointer. Will update this thread once I hear back.