AfterCare detects whether your PDF contains a real text layer or is a scanned image. Documents with selectable text are read directly, which is faster and more accurate. Scans and photos of pages — including images you take with your phone — go through a vision-based transcription step that converts the image to text before the remaining steps run. You do not need to do anything differently: AfterCare chooses the right path automatically based on what it finds in your file.
The eight steps
1
Reading the text on your document
AfterCare reads the raw text from your file. For PDFs with a real text layer, this step extracts the text directly. For scanned pages and photos, it uses vision-based transcription to turn the image into numbered lines of text.What you get: A line-numbered transcript of your document that every later stage can reference. If this step cannot produce a usable transcript, AfterCare stops and tells you — it will not try to build a recovery plan from nothing.
2
Sorting it into sections
AfterCare reads the full transcript and maps out the overall structure of the document: which sections deal with medications, which cover follow-up care, which list warning signs, and so on.What you get: A structured view of your document that guides the five steps below. Like the reading step, this one is essential — if the document cannot be structured at all, AfterCare stops and reports the problem rather than producing unreliable output.
3
Finding your medications
AfterCare locates every medication mentioned in your document and extracts the name, dose, frequency, timing, and any instructions about food or missed doses.What you get: A medications list where each entry cites the exact line it came from. If your document has an unreadable or missing medications section, you get an empty medications list and a note about the gap — not a fabricated list and not a broken guide.
4
Finding your appointments
AfterCare scans your document for follow-up appointments, including the provider name, specialty, location, and date where the document gives them.What you get: An appointments list with one entry per follow-up. Each appointment cites its source lines. If the document mentions no appointments, the list is empty.
5
Finding your warning signs
AfterCare pulls out every warning sign or symptom the document lists, along with the action it calls for: contact your provider, go to the emergency room, or call 911.What you get: An emergency screen that presents each warning sign exactly as your document describes it, with the corresponding action. Nothing is added or softened.
6
Building your recovery timeline
AfterCare assembles a chronological timeline of your recovery from the instructions scattered across the document — rest periods, activity restrictions, wound-care steps, and any milestones the document marks.What you get: A day-by-day or week-by-week timeline so you can see where you are in your recovery at a glance.
7
Explaining the medical terms
AfterCare identifies medical terms and abbreviations in your document and writes a plain-language explanation for each one, sourced from the context your document provides.What you get: A terms glossary where every definition points back to the line in your document where the term appeared, so you can read the original alongside the explanation.
8
Double-checking every detail
A separate verification pass re-reads your document and checks every item produced by the five detection steps. It confirms that each item’s cited lines actually appear in your original document and that the citation genuinely supports the extracted information.Any item whose citations do not check out is dropped entirely. Any item that cannot be traced to a real line in your document is removed or flagged — it will never appear as settled fact in your guide.What you get: A recovery plan where every displayed item has been independently verified. Items that do not pass are either removed or flagged for you to check in the original document.
Each section of your guide is independent
The five detection steps — finding medications, appointments, warning signs, building the timeline, and explaining terms — run independently of one another. A problem with one section of your document affects only the corresponding part of your guide. If your document has a damaged or missing medications table, you receive an empty medications list and a note about the gap. Your appointments, warning signs, and timeline are unaffected and still appear in full. One bad section cannot take down the whole guide.Grounding and citations
Every item in your recovery plan carries two pieces of provenance: a set of source line numbers pointing to the exact lines in your original document, and a confidence score between 0 and 100.- Items with strong support from your document appear normally in your guide.
- Items with weaker support are flagged with a prompt to check the original document rather than being presented as settled.
- Items with no valid citations are dropped entirely during the verification step and never reach your guide.