Yes: https://github.com/beacondb/beacondb-unifiednlp

You dont need a second computer, just replace the drive with an empty one.

It does (they of course use their Azure Blob storage under the hood). Forgejo however does not (even though it supports it for releases and packages and Golang has this in their standard library).

I’ll keep the idea in mind.

Technically unlimited, but you obviously need to have enough storage to cache the zip files (and RAM to cache the file index). My server is very small, so I needed to set the limit low.

No, that temperature would damage your screen. The professional hot plates for phone repair are typically set to 85-90°C. With a heat gun you may need to set a higher temperature since you are only heating up part of the phone and it cools down again during the process. My printer (Prusa MK3) with PCB heater can go up to 120°C, so it looks perfect for the job.

Bad bot, you should verify whether Piped can accept an URL and not simply link all YouTube URLs to Piped.

RSS feeds are XML files which contain a list of documents hosted on the internet (articles, audio/video). The feed entries contain basic metadata (title, date, author, summary) and a link to the original website (or audio/video file in the case of a podcast).

Feed readers send a simple web request to the website hosting the feed, downloading it if it has changed since the last update. The content is then combined with other feeds and displayed. This way you can have a personalized news reading experience without needing to create an account at a a central provider or open every individual site.

Alternative YouTube clients use RSS feeds provided by YouTube (example: https://www.youtube.com/feeds/videos.xml?channel_id=UC2DjFE7Xf11URZqWBigcVOQ), but they are only used to update subscriptions. All other requests (search, watching videos) are handled by the same web interface as the YouTube desktop application. Fetching the RSS feeds is a lot faster than opening the channel page, so the RSS featuee allows you update 100 or more channels in a few seconds.

The way podcast ads work is either just like YouTube sponsorships (the podcaster gets paid by a company to speak an advertisement themselves) or they are dynamically inserted by the podcast provider (these are the interrupting ads). Since most podcast apps dont store cookies, there is no way to track users and personalization is done only via the IP-based location and topic of the podcast. RSS-based podcast players have no way of directly reporting back playback telemetry. The server hosting the podcasts can only count the number of downloads/playbacks. So there is no way to count the amount of watched ads when using a RSS-based podcast player like AntennaPod or Kasts. Note: this does not apply to podcasts on Spotify, Apple Music or similar platforms. These platforms absolutely track your listening activity. I have no idea whether this affects ad/sponsorship earnings.

One important thing if you are building a RSS application is that the server should support conditional requests (the If-Modified-Since header). This way, a client does not have to download the entire feed on every update. It simply sends the last update date with its request and the server returns an empty response if the feed is up to date.

There are some applications (for example YouTube) which dont support this, resulting in higher-than-necessery data usage, especially on mobile.

You also cannot use it to store secret information like bank account/credit card details, API keys, etc.

I think they covered it with black foam, because they do not want to reveal what camera they use.