Product updates

Workspace upgrades

• Your workspace now accepts only signed FlowKeyOps releases. An upgrade whose signature does not check out is refused, and nothing is changed.
• Cancelling a scheduled upgrade never leaves a workspace half-upgraded.

Reliability

• If building a new workspace fails part-way, FlowKeyOps now finds and removes everything that build had already created, and frees its place automatically, instead of leaving it for an operator.
• A workspace you have paid for is never released before the end of the period you paid for.

Billing

• Billing dates and times in your panel and billing e-mails are shown in your own time zone, with the zone named, instead of a server time.

Verify the agent before it runs

• From version 0.32.1, every FlowKeyOps agent release is signed. Your security team can check that the agent running in your network is the one FlowKeyOps built, before it runs.
• Each version is published openly, with no account needed: the signed container image, Linux binaries (amd64 and arm64), their SHA-256 checksums with a signature over the list, and a software bill of materials (SBOM).
• The new "Verify the FlowKeyOps agent" article (Secrets and Security) has our public key and the exact commands. Agent versions before 0.32.1 were not signed; use the update notice in your panel to move to a signed version.

Linux (systemd) installs

• The Linux (systemd) installer now downloads the agent binary for the version FlowKeyOps recommends, checks it against the published SHA-256 checksum, and refuses to install it if they do not match.

Agent 0.32.1

• Same agent behaviour as 0.32.0. It is the first signed release, and its image is now published under gitlab.flowkeyops.cloud. Agents on 0.32.0 will see an update notice.

Restoring your previous workspace

• When you subscribe again after cancelling, your previous workspace is restored automatically — and now the people who were members of it come back with it, up to your plan's seats. Each one is recorded in your Audit history.
• If your previous workspace holds more environments or members than your new plan includes, FlowKeyOps asks before restoring. A "Choose what to restore" card on the Backups page, pointed to from your Overview, lets the workspace owner or an admin pick what to bring back, up to your plan's limits. Nothing is restored until you choose, and your new workspace works normally in the meantime.
• What you leave out is not restored. It stays in the backup until the backup's retention period ends — the card shows the date — and can be brought back before then by moving to a plan that includes it and asking our support team for a restore. Earlier versions of left-out workflows remain in your repository's history.
• After the restore, the workspace owner receives an email listing anything that was left out and the date it can be recovered until.
• Restoring a previous workspace needs a paid plan.

New: Enterprise

• A new Enterprise plan for organizations running automation at scale: 25 users, up to 10 connected runtimes, two years of audit history, and hourly encrypted backups with 48 hourly and 90 daily restore points, in a private security cell dedicated to your organization.
• Subscribe from the Pricing page, monthly ($999) or annually ($9,990). After you subscribe, our operations team prepares your private security cell, and your control panel shows its progress.
• Single sign-on, SCIM provisioning, custom roles, multi-stage approvals, audit streaming, and priority support are on the Enterprise roadmap and are listed on the Pricing page as not yet available.

Team

• Team is now $299 per month or $2,990 per year for new subscriptions. Existing subscriptions keep the price they were sold at.
• Team now runs on shared security cells, so a new Team workspace is ready as quickly as an Individual one.
• You can add members beyond the five Team includes, at $20 per member per month (or $200 per year on an annual plan), from the Members page. Each purchase is its own subscription; cancel it under Billing to remove those seats at the end of the period you paid for. If your workspace ever has more members than seats, everyone keeps access, but new invitations wait until a seat is free.

Enterprise Custom

• The Custom plan is now called Enterprise Custom. It starts from Enterprise and is quoted for your organization.

Your time zone

• Every time in the control panel is now shown in your own time zone, with its abbreviation (for example EDT or JST), instead of UTC.
• FlowKeyOps detects your time zone from your browser the first time you open the control panel. To change it, go to Profile and choose your time zone under Time zone. It applies only to you: each member of your team sees their own.

Audit

• The Audit calendar now groups events by your own days, so something that happened at 9 pm your time appears on that day, not the next.
• Audit exports write each time in your time zone with its offset from UTC (for example 2026-10-12T20:50:00-04:00), so the file is still unambiguous for anyone who reads it.
• On the Audit calendar, the icon for control panel and access events now matches the icon on each entry in the list.

Your vault's recovery keys

• FlowKeyOps keeps a sealed copy of your workspace vault's recovery keys outside the cell your workspace runs on. Nothing in the control panel or the control plane can read it: only the FlowKeyOps owner can retrieve it, by hand, and our procedure requires a support request that cites a support ticket first.
• If FlowKeyOps support ever needs those keys to repair your vault, you can now see it on the Audit page of your control panel. "Support requested emergency vault access" shows the support ticket the request cites, and a second entry follows when it is closed: "Emergency vault access completed" or "Emergency vault access request cancelled". The keys themselves never appear anywhere in the control panel.

Credentials

• You can now create a credential in the control panel and have it pushed to your n8n. Give it a new name, enter its values and press Save and push: FlowKeyOps writes it to your own vault and creates it in your n8n. Before this, the control panel could only fill in credentials that already existed in n8n.
• To update a credential that already exists in your n8n, use its exact name, or open it from the credentials list.
• "Add a credential" now sits at the top of the Secrets page, and every message on that page appears in the same place, above it.

Your plan

• The control panel now shows the limits of your plan accurately. An Individual plan includes one environment and one user. Team includes three environments and five users.

Your n8n API key

• Replacing an environment's n8n API key now works every time. A second or later replacement on the same environment used to fail with "could not replace the key"; each one now reaches your running agent within a minute, with no reinstall.
• Once your agent has received your n8n API key, FlowKeyOps removes it from its records. If you later reinstall the agent or add another one to the same environment, it starts without the key: use Replace n8n API key and the key is delivered to it within a minute.

Messages in the control panel

• The messages on the Environments page are now colour-coded: blue while FlowKeyOps finishes the work, green when it is done, red when something was refused. Confirmations clear themselves after a few seconds, so a second action always shows a fresh message.

Seven weeks of changes since 1.0.0, grouped by what you can now do.

Health and insights

• A new Insights page answers "how is my automation changing, and what needs attention": change activity, sync health, promotion flow, credential hygiene, the applications your workflows use, and how often each workflow ran in n8n (counts only; turn it off in Settings).
• Health is the first Insights tab: whether each n8n environment and each part of your workspace passed our latest check, with uptime over 24 hours, 7 days and 30 days, recent incidents, your backup status, and what your workspace has reserved against your plan's allowance. Uptime is a measurement, not a service-level agreement, and a check we could not make never counts as downtime.

Your team

• Invite teammates by email, give each a role, and suspend, reinstate or resend an invitation. Invitations are limited to your company's email domain, and a teammate refused an action is told which role they need.
• You can grant FlowKeyOps support read-only access to your workspace for a limited time, and revoke it whenever you like. Support never sees a stored secret value, and every view appears in your audit history.

Backups and restore

• A Backups page lists every backup of your workspace, kept for your plan's retention period.
• You can ask for your workspace to be restored from a backup; you confirm by typing your workspace's name.
• If you cancel and subscribe again within 30 days, your previous workspace is restored into the new one automatically, including your connected agents.

Workflows and promotion

• Promotion paths between environments, with a check that the target has every credential your workflows need before anything moves.
• Undo a promotion, and restore any earlier version of a workflow — view it, and compare it with what is running, before you restore.
• Drift shows a side-by-side, colour-coded comparison with both remedies on the row, and tells you how recently it was checked.
• Each workflow's n8n tags, a Find box and sorting on the Workflows table, and an attention list on each row with a fix beside each problem.

Secrets

• Saving a credential pushes it to n8n, not only to your vault; you are asked first when workflows use it.
• Rename a credential in place, and n8n follows the rename.
• Deleting a credential also removes it from n8n, with live progress, and a credential a workflow still uses cannot be deleted.
• Each credential says which fields hold no value and how to fill them, and which workflows use it.
• Copy all secrets along a promotion path in one step.

Getting started and support

• A five-step setup wizard until your first agent connects, then a guided tour on every page.
• While your workspace is being built, a progress bar shows each step; if it is waiting for capacity, the panel says so rather than showing progress that is not happening.
• A Support page: open a ticket, follow the conversation and reply, and paste screenshots, without leaving the panel.
• The workspace owner is emailed if an agent has stopped reporting for 7 and then 14 days.
• Every button now shows a message the moment you press it.

Billing

• If a renewal payment fails, your workspace becomes read-only rather than being shut off, and returns to normal as soon as payment succeeds.

Known limits at this release

• Drift is found by scheduled checking, not notification: n8n does not announce an edit, so a direct change is seen at the next check.
• Credential VALUE differences cannot be detected on any automation platform, because none allows a stored value to be read back.
• A restore request is carried out by FlowKeyOps staff, not instantly.
• Resources shows what your workspace has reserved, not live usage.

Establishes a version for a service that had 161 commits and no tag.

The number is new; the software is what has been running.

Customer-visible capability at this tag:

• Workflow inventory, Git-backed approved versions, and drift detection
• Resolve drift in either direction: restore the approved version into n8n, or accept the running version as approved
• Credential storage in a per-customer vault, with forms built from the customer's own n8n; missing and orphaned credential reporting
• Named environments and generated installers (Docker, Kubernetes, Docker Compose, Linux systemd)
• Append-only audit history with export and plan-based retention
• Automatic provisioning from a purchase; verified encrypted export before any data is removed

Known limits at this release:

• Drift is found by scheduled checking, not notification: n8n does not announce an edit, so a direct change is seen at the next check.
• Credential VALUE differences cannot be detected on any automation platform, because none allows a stored value to be read back.
• Changing plan between Individual and Team is not self-serve.

Full notes: CHANGELOG.md

• Where a live workflow no longer matches its approved version, see both sides side by side.
• Resolve the difference in either direction: restore the approved version into n8n, or accept the running version as the new approved one.
• Restoring content never switches a workflow on or off. Its enabled state is left exactly as you set it.
• Once you resolve a difference it stops being reported. Previously an accepted change still showed as outstanding even though it had worked.
• When an action fails, the panel shows what your n8n actually said rather than a bare error label.

• A credential you enter in the panel is delivered all the way into your n8n, sealed at every step.
• If your n8n refuses a credential you are told so, and told that retrying will not help, so you know the value itself needs changing. Previously such a credential said "still saving" indefinitely.
• Non-secret parts of a credential are shown when you edit it, so you only retype what you are actually changing.
• See how many of a credential's fields still need a value.
• The panel tells you when your connector is out of date and how to update it, and will never suggest moving to an older version than the one you run.

• Your activity history now shows what actually changed, and how much, instead of only that something changed.
• Opening the dashboard no longer fills your history with noise. Roughly six in ten recorded events were previously just page loads.
• See at a glance which kinds of activity your workspace has recorded.
• Download the connector for your environment as a single file whatever platform you chose, and the file name tells you which version you have.
• Be told when a workflow needs a credential your n8n does not have — one of the most common causes of a workflow that fails for no visible reason.

• FlowKeyOps finds the credentials your workflows reference, so you do not have to list them yourself.
• Credential forms show exactly the fields your n8n shows — same labels, same hints, same help text — instead of a generic form.
• The form matches the n8n version you actually run, including credential types that inherit their fields from another type.
• Where a field has a sensible default it is offered as a suggestion, so there is less to type.
• See stored credentials that nothing in your n8n uses any more, so you can clean them up.

• Create named environments such as Development and Production, and manage each one separately.
• Get a ready-to-run installer for the environment you are setting up: Docker, Kubernetes, Docker Compose, or Linux systemd.
• Store credentials so that only your own environment can reach them. A credential created in Production is invisible from Development.
• Credential values are sealed before they leave the panel and can only be unsealed by your own environment.
• Run more than one connector for the same environment, so a single machine going down does not disconnect you.
• Retire a connector you no longer use, and it immediately loses access.
• Rename a connector or adjust how often it checks in.

• Connect your own n8n by installing a small connector inside your own network. It reaches out to us, so you open no inbound access.
• What you install is a one-time, short-lived joining code rather than a lasting key: once used it is worthless, so a code left in a ticket or a chat cannot be reused.
• See every workflow in your n8n listed in the control panel.
• Keep an approved version of each workflow, so there is always a known-good copy to compare against.
• Your workflow list refreshes on a schedule set by your plan, and you can also refresh it yourself on demand.
• If your n8n is unavailable, the panel says so plainly rather than silently showing stale information.

• When a subscription ends, an encrypted copy of your data is produced first — and checked that it can actually be read back — before anything is removed.
• That copy is written to two independent places and both are verified, so a single storage failure cannot lose it.
• Your copy is kept for 30 days after your subscription ends, then removed automatically.
• Support can tell you the status of that copy while it is being prepared.
• Your workspace is built and confirmed working before it is marked ready, so you are never handed an environment that is still being assembled.

• Buying a subscription on the website now sets up your workspace automatically, with no wait for someone to do it by hand.
• If a payment confirmation is delivered more than once, your account is only ever changed once.
• Your account keeps a permanent, tamper-evident record of every change made to it.
• How long that history is kept is set by your plan rather than by a fixed rule.
• Cancelling keeps your access until the end of the period you have already paid for, instead of cutting off immediately.

• Your workspace now has a permanent identity of its own, so your data stays yours even as the service grows underneath it.
• The control panel shows your real subscription and workspace details instead of sample values.
• Your workspace is given capacity automatically when you sign up, rather than being set up by hand.
• Capacity another customer has finished with is only reused once it has been confirmed genuinely empty — never on assumption.
• Sign-ins to the panel are short-lived and limited to your own workspace; an attempt to read someone else's is refused.
• A background health check runs continuously and flags anything that looks wrong before it affects you.