An archive your team
can work with.

01 / RECORDS AND RETRIEVAL

Find records across sources.

Search permitted archives together or select one source. Exact record and encounter identifiers retain their original formatting. Names use prefix matching. Additional filters cover birth date, document type and document dates.

• Browse by newest or oldest document date, with pagination and shareable search links.
• Open originals in the embedded viewer or a new tab, download files, and prepare a combined print PDF.
• Download selected files or all matching results as a ZIP within configured document and byte limits.
• Export manifests include source identifiers, sizes and file hashes.
• Send server ZIP jobs to approved Azure Blob or Windows share destinations. My exports shows progress and results after the browser closes.

Partial results are identified. Complete matching-result packages are refused when a source failed or configured limits are exceeded.

02 / SOURCES AND STORAGE

Keep the storage you choose.

Each archive has its own source definition, mapped columns and document root. Source connections are read-only. Administrators discover columns, map required fields and validate the mapping before enabling search.

Direct lake-file ingestion and arbitrary CSV or Parquet searching are separate future connectors. Schema changes require mapping review. Database-provider support does not establish support for every vendor’s exit format.

03 / ACCESS AND EVIDENCE

Decide who can do what.

Built-in roles cover records viewing, operations, export management, support and database administration. Custom roles assign feature grants and denials to directory groups or provider-scoped users. Archive permissions are evaluated separately from feature permissions.

Archive restrictions apply before search queries and on direct document routes, exports and dashboard views. Navigation follows effective permissions, while the server enforces access independently.

Current activity and original history

Access history records searches, views, downloads, print preparation and export outcomes. Records link to their history. Authorized users can filter by account, archive and date, export events or prepare patient audit reports.

The optional historical audit module preserves original source files in customer-configured Azure Cold storage and indexes explicitly mapped events from CSV, JSON or XML. Unknown formats can be preserved without inventing events. Original evidence remains separate from current ATLAS activity.

Integrity chains support verification. The organization defines retention, backups and independent checkpoints. An integrity chain does not make storage immutable against administrators controlling the databases and keys.

04 / OPTIONAL LOCAL AI SEARCH

Interpret a request on your server.

The included local model runs on the application server’s CPU and interprets short natural-language search requests into search fields. There is no external AI API, subscription or GPU requirement for this feature.

Users can review and correct proposed fields. Ordinary indexed search remains available when the model is disabled, busy or unavailable. Interpretation can be wrong, so explicit filters and user review matter.

The model does not read medical documents, create clinical summaries or generate SQL. AI search requires its own feature permission and ordinary search permission. Subsequent record access retains the same archive restrictions.

05 / EPIC CHART LAUNCH

Open the mapped patient archive.

The implemented SMART on FHIR EHR-launch connector opens mapped archive records for the selected patient. Optional Epic OpenID sign-in uses approved identities and explicit roles. Exact identifier systems and reviewed crosswalks establish patient mappings.

Customer acceptance requires a registered client, approved endpoints, patient mappings and non-production workflow testing. This does not establish Epic certification. Epic ingestion, document write-back and embedded Hyperdrive workflows are separate future work.

06 / OPERATIONS AND LICENSING

Give IT a view of the archive.

• Dashboard charts show indexed counts and date coverage, with searches for selected periods.
• Document verification checks file availability and identifies incomplete scans.
• Database checks and source-specific index plans help DBAs assess performance.
• Status, application logs and diagnostics show dependencies and configured peer responsiveness to authorized users.
• Configuration replication supports multiple nodes with explicit coordination and saved-versus-running warnings.
• Account preferences include theme, chart palette, viewer choice and time zone.

Patient-capacity licensing

Software licensing is based on indexed patient capacity rather than document count or web-node count. Signed licenses, usage counts and expiry information are available to authorized licensing administrators. Your order defines fees, term, scope, support and service levels.

Multiple nodes need suitable storage, identity, load-balancer and audit-database setup. Configuration replication does not replicate archive contents or guarantee uninterrupted jobs. Review those dependencies during deployment planning.

07 / AVAILABILITY AND ROADMAP

Current features and work ahead.

The public site illustrates the product. It does not connect to patient data or provide a live archive service.