Frequently asked questions

The business problem in plain language.

What does Liulum do?+

Liulum monitors a critical business path from customer intent to outcome: for example storefront, checkout, payment event, automation, and fulfilment. It connects related failures into one evidence-backed incident and, on eligible customer-owned infrastructure, can run only the recovery actions your policy permits.

How is that different from uptime monitoring?+

A homepage can return 200 OK while checkout JavaScript is broken, a payment webhook is missing, or a workflow reports success without creating the order. Liulum checks configured handoffs and expected outcomes instead of treating one green URL as proof that revenue is flowing.

Who is it for?+

SaaS and ecommerce operators, and agencies responsible for client systems. Current platform scopes include Shopify observation, WooCommerce and headless stacks, APIs, deployments, and n8n, Make, Zapier, Python, or custom workflow signals when the required connection exists.

Does connecting an integration mean it is healthy?+

No. A connection starts as unknown. Liulum marks it connected or healthy only after the configured test records evidence. If a provider cannot be checked, the portal should show unavailable instead of inventing a green state.

Can Liulum restart a failed n8n, Make, Zapier, or Python workflow?+

Monitoring can use signed events, heartbeats, callbacks, and reconciliation. A retry or replay is available only when the provider exposes a supported control, the operation is idempotent, the customer has authorized it, and the plan and policy permit it. Otherwise Liulum alerts or proposes the action.

What can Recover do automatically?+

On eligible customer-owned infrastructure, Recover can use pre-approved, reversible playbooks such as rollback, redeploy, restart, cache clearing, or provider-supported replay. It independently verifies the expected outcome afterward. Shopify-managed infrastructure is not repairable by Liulum.

What will Liulum never change blindly?+

Payment movement, secrets, destructive database operations, unsupported third-party infrastructure, and unapproved code changes. A high-impact or uncertain action stops for human approval.

Can I test it before paying?+

Yes. The Guided Proof is manually approved, lasts seven days, covers one real project and one revenue path, requires no card, and keeps autonomous recovery off. You receive the evidence actually collected.

What are the India plan prices?+

Watch is ₹2999/project; Continuity is ₹8999/project; Recover is ₹19999/project for an eligible project; Grid is ₹39999/month for agencies. Taxes and any metered add-ons must be shown before payment. See /pricing for USD prices and exact limits.

Which plan includes autonomous recovery?+

Watch is observe-only. Continuity adds deeper paths, automation evidence, reconciliation, AI-assisted explanation, and approval-based repair proposals. Recover is the plan that can allow guarded autonomous recovery after technical eligibility and policy setup. Grid does not automatically grant Recover to every agency project.

Can I get a surprise charge?+

No extra plan, add-on, unusual provider usage, or bespoke work should be charged unless its price is shown before payment or you approve it in writing.

What is the fastest way to start?+

Request the Guided Proof at /start. We confirm the project is a fit, create your client access, connect one real revenue path, and test each required connection before monitoring begins.

What access do you need?+

Observe-only checks may need only a URL, health endpoint, or signed event. Deeper telemetry and recovery require the smallest scoped, revocable credential the provider supports. The required access depends on the platform and enabled feature.

Where do clients and staff sign in?+

Clients enter the client-only dashboard and portal. Internal staff use the separate admin area. Server-side role checks block one role from entering the other role's application shell.

Does AI control my infrastructure?+

No. Deterministic checks decide permissions, scope, risk class, idempotency, budget, and approval. AI may help explain evidence or draft a proposal, but it does not bypass those controls.

Is customer data separated?+

Application queries are tenant-scoped and the database has row-level security policies as a second boundary. Privileged cross-tenant operations are reserved for audited internal jobs and administration.

What happens if Liulum or an AI provider is unavailable?+

Deterministic monitoring, durable event receipts, retries, dead-letter handling, and existing playbooks do not depend on an AI answer. If a required provider or database is unavailable, the system should fail closed, record the failure where possible, and avoid claiming a repair succeeded.

Can I see what happened?+

The portal uses recorded checks, incidents, operations, and reports. Missing evidence is shown as unavailable. Every allowed recovery should record its authorization, action, result, and independent verification.

Who do I contact?+

Email [email protected] or use the customer-care assistant in your client portal. The assistant can use your tenant-scoped account context, but billing, legal, security, and consequential changes are escalated to a human.

How do I cancel or request a refund?+

Email [email protected] or use the available billing controls. Cancellation stops the next renewal after the current paid period. Refund eligibility is governed by /refunds and any signed order.

Email [email protected], read Support & customer care, or see full plan details on Pricing.