Searchable archive for every agreement you have signed

I want an app that turns every document we have ever had signed into a searchable archive I can actually read. Today the executed PDF is slow to find after the fact, and the values people typed into the form stay trapped inside the document, so answering a simple question about what we agreed means opening files one at a time. Build this as a retrieval and reading surface over our completed DocuSeal submissions, used mainly by legal, finance and ops.

The main screen is a browse view listing completed submissions from DocuSeal List Submissions, filtered to completed status. Each row shows the counterparty name, the template the document came from, the date it was signed, and whether it has already been abstracted or filed. Put a search box at the top that matches counterparty name using the submitter search parameter on List Submissions, plus a template filter populated from DocuSeal List Templates and a date signed range filter. Search is the priority here, so paginate the list, keep the filters visible on the same screen rather than behind a menu, and make the whole thing feel fast on a large archive.

Clicking a row opens a detail page that puts the executed document next to the data behind it. On one side, render the final signed PDF fetched with DocuSeal Get Submission Documents. On the other, list the field values each signer actually submitted, read with DocuSeal List Submitters scoped to that submission and DocuSeal Get Submitter for each signer, which returns the submitted field values. Show every field as a label and value pair grouped by the signer it belongs to, so someone can read exactly what a counterparty entered without opening the PDF at all. Use DocuSeal Get Submission for the header details: document name, completion date and signer status.

Put an "Abstract this agreement" button on the detail page that starts a background agent. The agent reads the signed document and the captured field values for that submission, then extracts the parties, the effective date, the term length, the contract value, the notice window and the key obligations. It appends that as one structured row to a Google Sheets contract register using Google Sheets Append Values, with a column per extracted term plus counterparty, template, date signed and a link back to the submission. This part matters: whenever the agent cannot find one of those terms in the document, it must write an explicit "not found in document" marker for that field instead of guessing or leaving the cell silently blank, and the app must display that marker clearly so a reader can tell a missing clause from an unread one.

Add a second tab in the app that shows the contract register itself, read with Google Sheets Get Values against the register sheet. It is a table of every agreement that has been abstracted, with the extracted terms as columns, searchable the same way the browse screen is. When the background agent finishes, refresh this tab so the new row appears without the user reloading, and link each register row back to its detail page.

Add a "File a copy" action on the detail page that builds a clean Drive archive alongside the register. It creates a folder named after the counterparty with Google Drive Create Folder, reusing the existing folder when one is already there by checking with Google Drive List Files first, then stores the executed PDF inside it with Google Drive Upload File (Multipart). Once a copy is filed, show the resulting folder link on the detail page so legal and finance can jump straight to the archived document.

Keep the scope tight. The priority is fast search across counterparties and showing the extracted terms clearly, including the explicit note on any field the agent could not find. This is a retrieval and reading surface, not a reminder system, so no scheduled digests, no chat alerts and nothing that runs on a timer. Renewal timing is simply one of the extracted fields here, not the purpose of the app.