A HIPAA-conscious CSV viewer that lets you view PHI without uploading it

For healthcare analysts, clinicians and clinical data teams, the riskiest moment in working with a patient export is rarely the analysis itself — it is the upload. The instant a file of protected health information (PHI) is dragged into a cloud spreadsheet or a web-based data tool, that data crosses the network, lands on someone else's server, and becomes another disclosure to document, another vendor to trust, and another surface for a breach. OmniGrid removes that step entirely. Your CSV is parsed and queried locally by a Rust engine compiled to WebAssembly, and the data itself never touches the network — there are zero network requests for the file you open.

Why local-only processing fits HIPAA-conscious workflows

HIPAA's Security Rule is built around safeguarding electronic PHI: limiting access, minimizing disclosure, and controlling how data is transmitted and stored. A viewer that never transmits your data supports those goals by architecture rather than by promise:

• No transmission of PHI. Because the CSV is read directly from your own machine, opening and querying it does not send PHI anywhere. There is no upload to intercept and no cloud copy to secure.
• No third-party disclosure for the file you open. OmniGrid never receives your data, so it is not acting on your PHI — which is why, for the file contents, there is generally no Business Associate Agreement to negotiate.
• Minimum necessary by default. You do not have to copy an entire patient export into a SaaS warehouse just to answer one question. The data stays exactly where it already lives.
• Confidentiality by design. With no upload endpoint and no server-side storage of your file, there is no bucket to misconfigure and no vendor breach that can expose the rows you were reviewing.

Let us be precise about what this does and does not mean. OmniGrid is a tool thatenables HIPAA-conscious workflows; it is not itself a certified covered entity, a business associate, or a compliance product. It removes the transmission and third-party-storage risk for your dataset, but whether any particular activity is HIPAA-compliant still depends on your full environment — your device security, disk encryption, screen-lock and access policies, and your organization's rules. Treat local-only viewing as a strong technical control that shrinks your risk surface, not as legal advice.

How it works under the hood

Drop a CSV into OmniGrid and the WebAssembly engine reads it directly from your disk. Heavy parsing runs inside a Web Worker, so the interface never freezes even on very large exports. Files beyond memory — 10 GB and up — are streamed in chunks into your browser's IndexedDB storage, which lives on your own machine, never a remote server. Everything runs in any modern browser (Chrome, Firefox, Safari, Edge), with no install and no account required to begin. When you are done, clear site data and the file is gone, with no residual cloud copy left behind.

On the free tier you can open a CSV up to 50 MB and view up to 10,000 rows — enough to sanity-check a de-identification pass or spot-audit an export before it moves. Upgrading to OmniGrid Pro unlocks unlimited file sizes, unlimited row display, a full in-browser SQL query engine with aggregates, joins and window functions, support for Excel, JSON, Parquet and log files, and data export to CSV or JSON — all still running locally. For a clinical analyst, the ability to run real SQL over a PHI export, and export a de-identified subset, without that export ever leaving the laptop is precisely the point.

Local viewing versus cloud spreadsheets and web tools

The practical difference is stark when you compare the two workflows on the same task: reviewing a spreadsheet of patient records.

• Cloud spreadsheet or web tool: upload the file → PHI crosses the network → stored on the vendor's servers → the vendor becomes a party to your data → you document the disclosure, chase a BAA, and trust their breach controls.
• OmniGrid: open the file → the CSV is read locally by WASM → filter and query it in the browser → export a de-identified result if you choose → the original PHI never left your device.

You get the same analytical power — filtering, sorting, joining, aggregating, exporting — with a dramatically smaller compliance footprint. This is the same privacy-by-design reasoning behind our other regulated-data guides. If your constraints come from EU data protection rather than US healthcare, the same never-leaves-the-browser approach is covered in ourGDPR-compliant data analysis guide, and you can browse the full set on the use cases page.

Good fits for healthcare teams handling PHI

1. Reviewing a patient or claims export before a migration, merge or deletion request.
2. Spot-checking a de-identification pass without piping PHI through a third party.
3. Auditing an EHR or lab-system CSV export for data-quality issues.
4. Ad-hoc SQL over an encounter list to segment cohorts or verify counts.
5. Inspecting log files that may contain patient identifiers or MRNs before sharing.

You can try the entire workflow right now — open OmniGrid in your browser and drop in a CSV. Nothing uploads, so there is nothing to undo and no PHI to claw back from a server.

Frequently asked questions

Is OmniGrid a HIPAA-compliant CSV viewer?

OmniGrid is designed to support HIPAA-conscious workflows: it opens and processes your CSV file entirely inside your own browser using a Rust/WebAssembly engine, so protected health information (PHI) is never uploaded to our servers or any third party. That local-only architecture removes the transmission and cloud-storage risks that usually create HIPAA exposure. To be precise, OmniGrid is a tool, not a certified covered entity or business associate — actual compliance depends on your entire environment, including your device, disk encryption, and organizational policies.

Does opening a file with PHI count as disclosing it to OmniGrid?

No. Because parsing and querying happen locally in your browser, the contents of your file are never sent to us. There is no upload endpoint and no server-side handling of the data, so OmniGrid does not receive, store, or transmit your PHI. This is why, for the data you open, there is typically no Business Associate Agreement (BAA) to sign — we are not acting on your PHI in the first place.

Do I need to sign a Business Associate Agreement (BAA) to use it?

For the file contents you view, there is nothing for us to process, so a BAA covering that data is generally not applicable — the PHI never reaches us. If you upgrade to Pro, your billing and account details are handled by our payment provider, but your dataset is never part of that flow. Always confirm the specifics against your own compliance program and legal counsel.

Where is my PHI stored while I work with a large healthcare file?

Small files are held in memory. Very large files (10 GB and up) are streamed in chunks into your browser IndexedDB storage, which lives on your own disk — never a remote server. No chunk is transmitted over the network. When you are finished, you can clear it instantly by clearing site data in your browser, leaving no residual copy on any cloud service.

Can I run SQL queries or export filtered PHI without uploading anything?

Yes. On the Pro tier, a full in-browser SQL query engine (aggregates, joins, and window functions) runs locally over your file, and you can export filtered results to CSV or JSON. Every one of these operations executes on your own machine, so you can de-identify, segment, or audit PHI without a single row ever leaving your device.