StackJack 0.1.0.8701+73f763d

This release makes moving your organization to another region more reliable and keeps you better informed while it happens. It also fixes an error on the Foundry setup page, makes automation decision steps more reliable, and improves the answers from the Discord support bot's documentation search.

Improvements

• Earlier region-move completion email. The email confirming a finished region move, which contains your new MCP address, now arrives 30 to 90 minutes after the move is verified. Previously it came about a day later.
• Failed region-move notice. If a region move fails, the organization owner now gets one email saying so. If you retry right away and the retry succeeds, you no longer receive a failure email for the earlier attempt.
• Pause after repeated failures. After two identical failures within the same day, self-service retries are paused and the Region page explains the next step. The message no longer says that our team has been alerted.
• Access changes are held during a move. While an organization is being moved, or while a copy of it is inactive, changes that would be lost in the new region are refused with a clear explanation. This covers:
• Adding members, sending invites, and promoting a member to Administrator
• Creating MCP clients and adding tool roles
• Transferring ownership and granting agent-runner access
• Turning a person's client access back on
• Creating members through directory sync

Removing access stays available, including demoting an Administrator, removing a role, and revoking access.

• Clearer Discord bot answers. Documentation search in the Discord support bot picks the right page much more often, especially for tool-reference pages. It now says "No page answers that directly" only when it is confident. Otherwise it shows the closest pages.

Bug fixes

• Region moves stopped partway for some organizations. Moves no longer fail after all data has been copied when the organization has any of the following:
• Credentials enrolled by an automation
• Older inactive client authorizations left behind after a person's sign-in identity was relinked
• Historical audit or terms-acceptance records from a relinked person
• Unusable invitations during a move. Inviting someone while a move was in progress could send a setup email for an invitation that could never be used. The invite is now refused before anything is created or sent.
• Unclear agent-runner access error. Granting agent-runner access during a move showed a generic error code. It now shows the actual reason.
• Repeated failures paused retries too eagerly. A single old failure combined with a new identical one no longer pauses self-service retries. Both failures must fall within the last day.
• Late failure emails. A failure email is no longer sent a day late after a failed move.
• Error on the Foundry setup page. The setup page could return an error when the Owner-rights check on your Azure subscription did not finish in time. It now shows a message that nothing was created and that you can check the subscription again. It also no longer wrongly says you lack Owner rights in that case.
• Decision steps in automations. An unusual reply from the AI provider could make a decision step fail outright instead of following the outage behavior you configured. The step now falls back as intended.

UI & UX changes

• Wording about the old copy. The region-move confirmation email, confirmation page, and Region page now say your old copy is removed after the move is verified. They no longer promise a fixed one-day cleanup or a manual review by our team.
• Inactive copies are labeled. An inactive copy left in a region after a failed move now says it is inactive and tells you where your live organization is.