i had nearly two thousand saved posts across twitter reddit pinterest and instagram that i wanted to import into mymind. the official api has a credit system that limits you to about five hundred urls per month. at that rate it would take over three months to import everything. the browser extension can save items instantly without hitting any limits so clearly there was a different path.
someone reverse engineered the internal web endpoints that the browser extension uses. instead of calling the public api with an api key and consuming credits the extension calls post access dot mymind dot com slash objects with browser cookies for authentication. no credits no rate limits same endpoint. i grabbed my cookies from the browser dev tools tested a single request and it returned 201 created immediately. from there it was just a python script looping through all the urls at one request per second.
the reddit and twitter posts imported fine as webpage cards. pinterest was trickier. the server side fetcher tried to screenshot each pin page but pinterest blocks server fetchers so every card showed a blank login page. the fix was using pinterest's public oembed api to get the direct cdn image url for each pin saving it as an image type instead of a webpage and then patching the source url back to the original pin so the card still links to pinterest. all in 1889 urls imported in under an hour with zero api credits consumed. the whole pipeline is now scripted with deduplication so i can drop export files into an inbox folder and only new items get imported.