2 ms·
1. API to extract and store metadata about an uploaded PDF (images, relevant field locations, signature boxes, etc.) 2. Tool that shows an image of each field
by twodave 7y ago
1. API to extract and store metadata about an uploaded PDF (images, relevant field locations, signature boxes, etc.)
2. Tool that shows an image of each field to the user and allows them to enter human readable prompts (title, description, input style).
3. A PWA that uses the prompts to help an end user complete the form, collects signatures, etc.
4. Add the ability to link/upload some tabular data and tie fields from the data source to fields on a PDF.
5. Communications—email a unique short link to somebody to request they complete a form, pre-filled with data.
6. Build an example web site that consumes 1-4 and shows off e.g. completing a legal contract. Notifications when a form is completed.
7. Multi-party forms
8. Start figuring out how to charge people and who to sell to, build more niche wrappers around the product, stand up an auth/billing service, etc.
- richajak 7y agoThat's cool. I was working on something similar, extracting tabular data from PDF for the last 2 weeks. Curious about your method, as the process (e.g. to detect the coordinate of the header/footer fields, etc) is very manual and specific for each PDF template. Do you happen to encounter the same challenge? If yes, how do solve it, use machine learning? heuristics? etc?
- twodave 7y agoWell for now it is only attempting to read FDF or XFA forms, for which the metadata is pretty readily accessible. The tool will probably include a way to manually add/remove fields as well, but detecting form fields from basically an image is a level of awesomeness I have not yet set as a goal haha.