Help Center & FAQ

Sevrel connects to supported storage providers through their APIs. Provider-connected source files remain in that system, while Sevrel may store derived extracted text, citation excerpts, chunks, embeddings, and caches needed for retrieval. Files uploaded directly to Sevrel are stored by Sevrel.

Organization-scoped tables use PostgreSQL Row-Level Security and application-layer authorization, with automated cross-organization tests. These layered controls are designed to prevent cross-tenant access; no system should be described as making every future defect impossible.

Sevrel uses an enterprise AI provider with tiered model routing. Simple queries use fast models, complex analysis uses deeper reasoning models. The right model is selected automatically for each query.

Check that the document is accessible through your connected storage provider, is a supported format (PDF, DOCX, XLSX, text), and try rephrasing your query with specific property or tenant names.

PDF (text-based), Microsoft Word (.docx), Microsoft Excel (.xlsx), and plain text files — covering leases, rent rolls, OpEx budgets, and financial reports.

Sevrel works best with text-based PDFs. Image-only scanned PDFs without OCR may not be searchable. If your PDFs have been OCR-processed, the text layer is readable.

Sevrel uses retrieved source evidence when answering and is designed to include citations, but figures and citations can be incomplete or wrong. Always verify material figures against the source — Sevrel is a research accelerator, not a replacement for professional judgment.

Admins can invite team members from Settings. Invitees sign in with Microsoft Entra ID credentials. All access is invitation-only.

Viewer (query and view), Member (full feature access + sharing), Admin (user management, invitations, audit logs).

Click the source citations to verify. The AI may have found a different document version. Try rephrasing your query with more specific property or tenant names.

AI processing uses encrypted HTTPS connections. Relevant query and document content may be sent to the AI provider configured for your organization. Retention, abuse-monitoring, and model-training terms depend on that provider, account configuration, and governing contract; review Sevrel’s current Privacy Policy, DPA, and subprocessors list for the terms that apply. Each organization’s application data is scoped separately.

Sevrel’s architecture is designed for SOC 2 readiness with audit logging, access controls, and encryption. An independent SOC 2 examination has not been completed; readiness work is in progress.

Yes. Sevrel currently supports Egnyte, Google Drive, and direct upload. SharePoint / OneDrive and Dropbox are on the roadmap and cannot be connected today. Availability can depend on your organization’s configuration.

Sevrel can read provider files on demand and also stores derived retrieval data such as chunks and embeddings. Refresh and indexing behavior varies by provider and workflow, so a newly added or changed file may need a sync or re-index before semantic search reflects it.

Yes. The web interface is responsive. While optimized for desktop, you can query and review responses on mobile devices.

Open Settings → Integrations and connect Microsoft 365 or Gmail. Sevrel stores the resulting OAuth tokens encrypted; it does not store your account password.

Open the + menu in chat and choose Prompt shortcuts. A shortcut inserts one-turn guidance into an ordinary conversation; it is not a dedicated deterministic pipeline or continuous monitor.

Sevrel has three configurable proactive workflows: morning briefing, email triage, and lease expiry. When enabled and successfully processed, they work on bounded available inputs and can place structured findings in notifications or later briefings. The lease workflow checks recorded end dates inside 30 days; it does not discover notice clauses.

Supported workflows can generate downloadable PDF, Word, Excel, or PowerPoint files. API and report exports vary by feature and plan; the retired legacy dashboard does not provide scheduled email reports.

The legacy dashboard route has been retired. Use Portfolio for property and tenancy intelligence, and Work Orders for operational activity.

The standalone critical-date tracker is retired. When enabled, the lease-expiry agent checks confirmed portfolio records for lease end dates inside 30 days. It does not extract or monitor renewal, option, covenant, or notice deadlines; verify those obligations against the governing document.

Yes. Use the share control to create a read-only snapshot of supported message content and evidence. Attachments, completed action state, and source URLs are not included automatically; access also depends on the share settings and link state.

The historical saved-template panel and custom-template API are not mounted. Current catalog entries are one-turn prompt shortcuts: selecting one inserts guidance into an ordinary chat, which you can review and edit before sending. It does not invoke a separate deterministic workflow.

Yes, for organizations whose plan and feature settings enable API access. See the API Reference for the currently documented endpoints and authentication requirements.

Sevrel is an AI-powered document intelligence platform for commercial real estate, built by Romulus, Inc. It gives owner-operators and property managers source-linked answers about their leases, budgets, and tenant files, and helps manage tenant document intake — with detected PII redacted before document text reaches any AI model.

Sevrel is made by Romulus, Inc., a Delaware corporation headquartered in Sunrise, Florida. Sevrel is the product and brand name; Romulus, Inc. is the legal entity. The only official Sevrel domains are www.sevrel.com and api.sevrel.com.

Yes. Sevrel is a commercial product of Romulus, Inc., a Delaware corporation with a published mailing address at 14101 NW 4th St, Sunrise, FL 33325. Its only official domains are www.sevrel.com and api.sevrel.com, and the team can be reached at [email protected].

No. Sevrel accesses AI models through its provider’s commercial API, under whose terms customer data is not used to train the models and retention is limited by default. Detected PII is replaced with typed [REDACTED_*] tokens before document text is embedded or sent to any AI model. The security overview at www.sevrel.com/security has more detail.

No. Sevrel is not a property management or accounting system and does not replace Yardi, MRI, or AppFolio, nor does it natively integrate with them. It works alongside the systems a team already runs, on the document layer — leases, budgets, and tenant files — where it provides source-linked answers.

Sevrel is priced per portfolio, scoped to your properties and usage. Pricing is sales-led — there are no self-serve tiers or published price points, and enterprise pricing is discussed after onboarding. To talk through pricing for your portfolio, request access at www.sevrel.com/request-access.

Onboarding centers on connecting an authorized provider: Egnyte or Google Drive for documents (direct upload is also supported), and Microsoft 365 or Google for email and calendar. Answers become available after provider discovery and indexing complete; timing varies with provider availability, file type, and queue state, so Sevrel does not publish a fixed onboarding timeline.