Secure, browser-based tools for pharmacovigilance, regulatory and medical affairs. All processing runs locally in the browser: no uploads, no accounts, no servers.
PraxDocs · PDF Compare · HighLit AI · HighLit · PatentCliff · FOCUS
Work in pharmacovigilance, regulatory affairs and medical affairs depends on dense documents: clinical study reports, SmPCs, patient information leaflets and regulatory submissions. Available tools are often costly, generic, or require confidential material to be uploaded to external servers.
These tools were developed by Dr. Prashant Patel, MBBS, MD, to address specific gaps encountered in practice. Each one is a single web page that runs entirely in the browser, with no installation and no data leaving the user's device.
Privacy by default. Regulatory and clinical documents are confidential. Tools that require uploading them create avoidable risk, so everything here is processed locally within the browser.
No installation. No software installation, licences or IT onboarding. Each tool is a single page that can be bookmarked and used immediately.
Built from practice. Features reflect real working requirements, such as column-aware comparison for multi-column SmPCs, and prioritisation shaped by how medical and regulatory deliverables are managed.
A complete offline PDF toolkit. Open praxdocs.html in your browser.
What it does:
- Merge — combine multiple PDFs into one; select specific pages from each
- Stamp / Watermark — overlay a PDF (e.g. a "DRAFT" stamp or a company logo) onto another
- Rotate — rotate pages individually or in bulk
- Convert to JPEG — export individual pages as images
- Compress — reduce file size with selectable quality presets
How to use:
- Open
praxdocs.htmlin Chrome or Edge - Select a tool from the sidebar
- Drag and drop your PDF files into the file area, or click to browse
- Configure options (page ranges, quality, etc.) and click the action button
- The processed file downloads automatically — nothing is stored anywhere
All processing happens in-browser using pdf-lib and PDF.js. No server involved.
A diff tool built specifically for pharmaceutical documents. Open pdfCompare.html in your browser.
What it does:
- Extracts text from two PDFs and compares them side-by-side or in a unified view
- Auto-detects two-column layouts — reads the left column fully before the right, which is critical for SmPCs and patient leaflets that use multi-column formatting (most generic diff tools get this wrong)
- Three diff granularities: line, word, and character level
- Similarity score gives a quick read on how much has changed
- Myers diff algorithm — fast even on large regulatory documents
How to use:
- Open
pdfCompare.htmlin your browser - Drop your original document into the left panel and the revised document into the right
- Choose your diff granularity (word-level is usually the most readable)
- Click Compare — additions are highlighted green, deletions red
- Toggle between unified and side-by-side views as needed
- Use Reset to start a new comparison
Tip for SmPCs and PIL comparisons: If your document uses a two-column layout, make sure the auto-detect is working — the column order matters for a meaningful comparison.
An AI-powered analyser for pharmacovigilance and medical affairs literature review. Open HighLit_AI.html in your browser.
Built for situations where keyword scanning isn't enough — when you need to understand context, extract structured safety signals, or produce PBRER-ready narratives from a stack of abstracts.
What it does:
- Smart Highlight — AI reads your abstracts and colour-codes safety signals by type without pre-defined keyword lists
- Signal Extraction — returns a structured list of PV signals with severity classification (red / amber / green) and the source text
- PBRER Summary — generates a narrative paragraph suitable for dropping into a PSUR or PBRER literature section
- Imputability — structured causality assessment applied to the submitted text
How to use:
-
Open
HighLit_AI.htmlin your browser -
Select your AI provider and enter your API key — stored only in your browser session, never sent anywhere except the selected provider
Provider Default model Anthropic claude-sonnet-4-6 OpenAI gpt-4o Gemini gemini-2.0-flash OpenRouter anthropic/claude-sonnet-4-5 -
Paste your PubMed abstracts, MEDLINE export, or any safety text into the input panel — multiple abstracts are supported
-
Optionally add product context (e.g. drug class, indication) for more targeted analysis
-
Select an analysis mode and click Analyse
-
Review the output in the results panel
Privacy note: Your text is sent to the AI provider you select, using your own API key. It is not stored or processed by any other server. Choose a provider consistent with your organisation's data governance policy.
Research use only: This tool has not been validated under GAMP 5 or equivalent computer system validation (CSV) guidelines. Use in a regulated GxP or pharmacovigilance environment without prior formal validation may not comply with applicable quality system requirements.
A browser-based keyword highlighter for safety literature. Open HighLit.html in your browser.
Built for pharmacovigilance and medical affairs workflows where you need to scan abstracts for safety signals, mechanisms, or population characteristics — and want to see them colour-coded at a glance rather than Ctrl+F-ing the same terms repeatedly.
Four input modes — switch freely between them:
- Single / Full-Text (default): Paste or type any text directly into the text box — a single abstract, a full-text article, or any free text. No file needed — open the tool and start immediately. The downloaded
.docxcontains the full input with highlights throughout; the in-browser preview is capped at 1,000 lines for performance but the output file is never truncated. - Batch (.txt): Upload any PubMed
.txtexport (Abstract format, MEDLINE/tagged format, or plain text) to process multiple abstracts at once. - Embase (.xlsx): Upload an Embase Excel export ("MS Excel — Fields by Column" format). The tool auto-detects the Abstract and Title columns; if the Abstract column cannot be identified automatically, a dropdown lists all column headers to pick from. Each article is output with its title as a bold underlined header above the highlighted abstract text.
- PDF (.pdf): Upload any PDF with selectable text — full-text journal articles, clinical study reports, or regulatory documents. The tool extracts text respecting 2-column layouts (standard in most journals) and outputs a highlighted
.docx. A page range selector lets you restrict extraction to specific pages, so you can skip journal headers on page 1 or the reference list at the end.
What it does:
- Keyword categories are pre-loaded and fully editable — one term per line, one column per category
- Each category gets a distinct highlight colour; a legend at the top identifies which colour means what
- Prefix stem matching — entering
injurcatches "injury", "injured", "injuries", and "injurious" without requiring exact terms - Matching is case-insensitive throughout
- Download the keyword table as an Excel file to edit offline, then re-upload the updated version
How to use:
- Open
HighLit.htmlin your browser - Select your input mode: Single / Full-Text (default) to paste text directly; Batch (.txt) for a PubMed export; Embase (.xlsx) for an Embase Excel export (use "MS Excel — Fields by Column" format); or PDF (.pdf) to upload a PDF article — adjust the page range after loading to exclude headers or references
- Review the keyword table — edit categories directly in the browser, or download as Excel, modify, and re-upload
- Click Process File — the preview renders the text with matching terms highlighted by category colour
- Click Download to save the highlighted
.docxoutput
No data leaves your browser. Works with PubMed, Embase, and PDF inputs — no database credentials required beyond your institution's normal access. PDF extraction requires a PDF with a selectable text layer; scanned image-only PDFs are not supported.
Research use only: This tool has not been validated under GAMP 5 or equivalent computer system validation (CSV) guidelines. Use in a regulated GxP or pharmacovigilance environment without prior formal validation may not comply with applicable quality system requirements.
Find which Orange Book–listed patents expire in a given date window. Open PatentCliff.html in your browser.
Built for portfolio, business development, and generic-entry planning, where the question is "which products lose patent protection in the next 12–36 months?" and the Orange Book data files are the starting point.
What it does:
- Reads the FDA Orange Book
products.txtandpatent.txtfiles and joins them on Appl_Type + Appl_No + Product_No - Date window — quick presets (next 6 / 12 / 24 / 36 months) or a custom from/to range
- Filters — RX, OTC, discontinued, RLD only, include delisted patents, and "only products whose last patent expires in range"
- Search by ingredient, brand name, or applicant
- Two views — product level (earliest and last expiry, patent count) and patent level (one row per patent, with substance/product flags and use codes); click any column header to sort
- Export — Excel workbook with Products, Patents, and Filters sheets, or CSV of the current view
How to use:
- Download the Orange Book data files from open.fda.gov or the FDA Orange Book page and unzip them
- Open
PatentCliff.htmland drop inproducts.txtandpatent.txt(any order) - Pick a date window and filters — results update instantly
- Switch between product and patent views, then download Excel or CSV
Download fresh files each time — FDA updates the Orange Book regularly. Excel export uses SheetJS, bundled in the page, so the tool works fully offline.
Research use only: Results reflect FDA Orange Book patent listings only. They do not account for regulatory exclusivities, patent term extensions, pediatric exclusivity, litigation, or unlisted patents. Not legal advice — verify before making business or regulatory decisions.
A daily planning tool built around three ideas: your most important tasks for the day, a time-blocked schedule, and a priority matrix. Open FOCUS.html in Chrome or Edge.
FOCUS is more involved to set up than the other tools because it stores data persistently. Read the setup steps below before opening it for the first time.
When you open FOCUS for the first time, you will see a setup screen before anything else. This is asking you where to save your data file — a plain .json file on your computer that FOCUS reads and writes to every time you make a change.
Step 1 — Choose where to save your data
Click Create New Data File. A file picker opens. Navigate to where you want to store your data and give the file a name (e.g. focus-data.json).
Recommended location: A cloud-synced folder such as OneDrive or Google Drive. This means your data is backed up automatically and accessible if you switch computers. A local Desktop folder works fine too — just be aware that clearing your browser data will not affect this file, but losing the computer will.
Step 2 — Grant permission
After choosing the location, your browser will ask you to confirm write access. Click Allow (or Save). FOCUS will create the file and load the main app.
That's it. You're in.
This is the most important thing to understand: every time you open FOCUS, you will see a reconnect screen. This is normal and expected — it is not an error.
Browsers require you to explicitly re-grant file access at the start of each session. This is a security feature of the File System Access API that cannot be bypassed.
What to do: Click Open My Data File → your browser will show a small permission prompt → click Allow. FOCUS loads in a second or two.
If you see "File not found" instead, it means the file was moved or deleted. Click Open Existing File and navigate to where your focus-data.json now lives.
| View | What it's for |
|---|---|
| Today | Your 3 Most Important Tasks for the day, plus your time-blocked schedule |
| Tasks | Full task list — add tasks, set due dates, assign categories, filter by status |
| Priority Matrix | Eisenhower-style quadrant view: Urgent/Important, Not Urgent/Important, etc. |
| Settings | Manage categories, change data file location, export/import data |
Most Important Tasks (MITs): The Today view shows three slots at the top. The discipline here is intentional — you are forced to decide, before looking at email or anything else, which three tasks will constitute a successful day. Mark them done as you go.
Time Blocks: Below the MITs is a time-block scheduler. Add blocks with a start time, end time, and label. This is your intended shape for the day — not a rigid schedule, but a plan that forces you to think about whether your MITs will actually fit.
Priority Matrix: Tasks you add anywhere in FOCUS can be dragged into one of four quadrants. This view is useful at end-of-week to see where your energy has been going versus where it should go.
Your data is saved automatically to the .json file on disk. To create an explicit backup:
- Go to Settings → click Export Backup
- This downloads a timestamped
.jsonfile you can store anywhere
To restore from a backup, use Import Backup in Settings and select the backup file. The app will reload with your imported data.
FOCUS uses the File System Access API, which is supported in Chrome, Edge, and other Chromium-based browsers, but not in Firefox or Safari as of 2025.
If you open FOCUS in Firefox, it will fall back to browser localStorage — your data is saved inside the browser itself rather than a file on disk. This works, but:
- Data is lost if you clear your browsing history or browser data
- Data does not transfer across browsers or machines
- There is no automatic file-level backup
If you use Firefox, use Export Backup regularly (Settings → Export Backup) and keep that file somewhere safe.
| Tool | Chrome | Edge | Firefox | Safari |
|---|---|---|---|---|
| PraxDocs | ✓ | ✓ | ✓ | ✓ |
| PDF Compare | ✓ | ✓ | ✓ | ✓ |
| HighLit | ✓ | ✓ | ✓ | ✓ |
| HighLit AI¹ | ✓ | ✓ | ✓ | ✓ |
| PatentCliff | ✓ | ✓ | ✓ | ✓ |
| FOCUS (file mode) | ✓ | ✓ | — | — |
| FOCUS (browser storage) | ✓ | ✓ | ✓ | ✓ |
¹ Requires an API key from your chosen provider (Anthropic, OpenAI, Gemini, or OpenRouter) and an internet connection.
Chrome or Edge is recommended for FOCUS to get persistent file-based storage.
No frameworks. No build step. No backend.
- JavaScript — all logic
- HTML5 / CSS3 — structure and a shared design system with light, dark and system-matched themes
- PDF.js — PDF rendering and text extraction
- pdf-lib — PDF creation and manipulation
- Myers diff algorithm — document comparison
- SheetJS — Excel export in PatentCliff
- File System Access API — persistent local file storage in FOCUS
- GitHub Pages — hosting
Developed by Dr. Prashant Patel, MBBS, MD, a medical doctor with more than nine years in the pharmaceutical industry across pharmacovigilance, regulatory affairs and medical affairs.
Feedback and issue reports are welcome via this repository's issue tracker.
Dr. Prashant Patel · LinkedIn · GitHub
MIT © Dr. Prashant Patel. Free to use, modify, and distribute.