Persephone is the fourth named version of Guara Cloud. If Odysseus was about navigating across the edges of your architecture, Persephone is about bringing more of the platform into your hands, with the same calm surface whether you are deploying one service, operating a full team, or asking an AI assistant to help.
This release moves in seven directions at once: enterprise control, a fuller service catalog, programmable operations, a home dashboard you can shape, cleaner investigations, safer deploys, and a public website that shows more of the real product. The result is a platform that feels broader without feeling heavier.
Enterprise Cloud
Accounts, SSO, contracts, members, and billing visibility in one place.
Guara Cloud now has a dedicated enterprise layer for teams that need more control. Enterprise accounts can manage members, invitations, access grants, contracts, plan details, SSO settings, and billing visibility without turning every request into a support thread. Admins get a purpose-built workbench for enrollment and lifecycle work, while users get a clearer enterprise dashboard that shows the account, contract, permissions, and plan context that matters to them.
Catalog 2.0
More managed services, clearer access, and lifecycle controls from the dashboard.
The service catalog is now much closer to a complete operating surface. Catalog services can show credentials and connection strings, expose selected public endpoints when a service supports it, list backups, and run lifecycle actions without asking you to leave the dashboard. The catalog also grew with Tailscale, WebhookX, n8n, and an Observability Stack that bundles Grafana, Prometheus, Loki, Tempo, and Alloy for teams that want their own monitoring stack beside their apps.
Guara Cloud, programmable
A new MCP server, a stronger CLI, and more ways to operate without clicking around.
Persephone opens Guara Cloud to AI tools through the new Guara MCP server, so assistants can use scoped platform tools instead of guessing from screenshots or stale notes. The CLI also grew into a much more complete operations companion: manage cron workers, inspect runs, trigger jobs, tail traces, render profiles, check vulnerabilities, and keep terminal workflows close to the same product model you see in the dashboard.
A home dashboard that is yours
Arrange the widgets that matter to your project, your service, and your day.
The Home page is now configurable. Add widgets for projects, services, deployments, logs, metrics, traces, profiling, topology, incidents, vulnerabilities, audit activity, and onboarding. Resize them, configure the ones tied to a project or service, and keep the signal you care about in the first view you open. It is still a dashboard, but now it starts with your mental model instead of ours.
Observability that reads cleaner
Logs, traces, metrics, and profiles with fewer raw edges.
Investigations got easier to read. Traces and topology now prefer service names you recognise over raw platform labels, platform work is grouped instead of scattered through your waterfall, and the profiling experience received a full visual pass with clearer flamegraphs, empty states, loading states, and service-detail integration. Logs, traces, metrics, and profiles feel more like one investigation path now, not four separate tools.
Safer deploys, clearer failures
New pushes take precedence, stuck states are rarer, and errors are easier to act on.
Deploys now handle real-world change with less drama. When a newer push arrives, the platform gives it priority instead of letting old work confuse the result. Build cancellations are easier to tell apart from real failures, rollout failures settle faster, and the dashboard translates more deployment problems into clear, localised messages with useful next steps. When something goes wrong, the screen should help you move, not make you decode platform jargon.
A clearer front door
The public site now shows more of the actual product, not a simplified promise.
The homepage hero was rebuilt around a guided product demo that moves through real Guara Cloud surfaces: projects, metrics, profiling, catalog, logs, and traces. The documentation grew with it, covering MCP, CLI cron workflows, service profiling, traces, catalog services, service networking, authentication, and apex redirects in both Portuguese and English. Persephone should be easier to understand before you sign in, not only after you deploy.
And across the platform
-
Magic-link login is now available, with a dedicated email template and clearer auth screens.
-
Build-cancelled notifications now read differently from genuine build failures.
-
Subscription cancellation now captures a reason, so billing conversations start with context.
-
Catalog service routes and public access settings are easier to understand from the service UI.
-
Docs search and docs links got sharper, including better result titles and cleaner entry paths.
-
The admin dashboard gained clearer financial snapshots, at-risk signals, and triage views.
-
More user-facing errors now use calm, actionable copy instead of exposing low-level details.
-
New catalog services show provisioning and readiness states more reliably while they start up.
-
Accessibility and bilingual copy were tightened across the new enterprise, catalog, home, and observability surfaces.
-
A quiet reliability pass reduced noisy alerts, improved backup coverage, and made platform health signals easier for the Guara team to trust.
And more, dozens of smaller cuts across the dashboard, CLI, docs, and platform edges, all aimed at making Guara Cloud feel wider, steadier, and easier to operate.