Monetize this tool - Githubissues

Benjamin-Loison commented 10 months ago

Could propose multiple services around this tool:

a SLA and ease in maintenance to empower various kinds of jobs such as community managers (cf #153), while no-key endpoint is debatable (however could make a web-scraping based adapter but supporting all features may be heavy in terms of workload and human work), just renting web-scraping only instances seem fair
add first and foremost client wanted features (cf #188)
propose hosting instances to not face CORS issues (cf #11) or just to reduce company or end-user workload (as it only retrieves the wanted data and not all the webpage for instance, it reduces computation and bandwidth needs, hence potentially lower latency even if it adds an intermediary)
donations (cf #177)
use a restrictive license but this quite unwanted (cf #94)
propose WebSocket/HTTP callback precising a given poll frequency (cf #71 and above 3.), it can reduce global workload as there is a single poll for possibly multiple clients

Related to #227.

Benjamin-Loison commented 7 months ago

Could highlight in #11 that renting a YouTube operational API instance solves the issue as well.

Benjamin-Loison commented 6 months ago

Note that web-scraping seems more stable than YouTube Data API v3, see this example. Especially as with web-scraping we can solve official API issues somehow without waiting a period of time that YouTube solves the issue or if they even at some point do so.

Benjamin-Loison commented 4 months ago

Should investigate private instances endpoint usage (web-scraping or no-key), but as #91 shows, the no-key endpoint may be faster but I think what people like is the unlimited quota, so somehow can leverage this to scale up the number of clients thanks to web-scraping.

Benjamin-Loison commented 1 month ago

Could provide a generic search engine with more parameters and results than YouTube Data API v3 and YouTube UI one.

To answer Stack Overflow questions like:

Both of these users do not seem to have private contacts.

https://discord.com/channels/933841502155706418/1018941529265553488/1288835857259696210

The searching algorithms can be open-source as the YouTube_captions_search_engine but the data dump may not be available. This might be charged as it requires a significant backend.

May can use an actual SQL database.

Could contact in private message the persons facing this need to propose this service.

Related to #4.

Could advertize all persons I redirected or dealt with my Stack Overflow answer 69259093, YouTube_captions_search_engine and #4.

When a search is being processed could do it on YouTube Data API v3 and verify that my algorithm treated the returned channels if not treat them before answering or specify that it is incomplete. Also if someone notice that would expect a given result, then can treat it.

I do not understand how crowd funding could help except if makes the service free, https://discord.com/channels/933841502155706418/1018941529265553488/1288852041103573084.

Maybe have legal issue with such a service even if it is free.

To have quickly an interesting database could first treat most subcribed channels pending. I suspect that it would be very effective.

Could name it QueryTube or DataTube, to keep the Tube part of YouTube, as You is too generic and Data from database.

I would find it interesting to establish such a database from web-scraping.

Using a free Oracle VPS at Marseille with 4 CPU entries in htop and 200 GB of storage only for this purpose looks like a good start for a very well thought approach.

A person:

``` -----BEGIN PGP MESSAGE----- hF4DTQa9Wom5MBgSAQdAeTV86HX2kpaXrAERwDvx6JuZ3QvJQC0s8OnPKhq1/Ecw 09uzQJyo3zPAd3MuQjkqvNe7Enc0v+LrhoLOZzP4wOJlroU4zUsG3b/mEsIDTkl/ 0kABpEloFmNMNnx90hKrX+x6R5Gj7SZgqtR+6tQge95Y1uoRsPKMx7Tu3fyWjeDp JhmustzYdIaOl8ipruEZg8Z0 =ChwS -----END PGP MESSAGE----- ```

does not know if can use some kind of enumeration for thumbnails for instance if have the default one or a given one. So enum of a binary or an id stating that it is the default one. Can just use null when it is the default avatar.

Benjamin-Loison commented 1 month ago

Searching Discord for 68970958 (from the Stack Overflow question 68970958):

https://discord.com/channels/933841502155706418/933841503103627316/1035959652397097021

Searching Discord for 69259093 (from the Stack Overflow answer 69259093):

Search Discord for YouTube_captions_search_engine (from YouTube_captions_search_engine):

https://discord.com/channels/933841502155706418/933841503103627316/1056948294015791115 https://discord.com/channels/933841502155706418/933841503103627316/1058400946838843422 https://discord.com/channels/933841502155706418/933841503103627316/1058489678925414455

Well too many entries to list them all.

For each person should consider the most appropriate message(s) and contact him about it.

How to code over Discord search? Maybe with DiscordChatExporter. To have a pretty link to click on with who (with what username) sent this message and if still on the Discord or can be contacted.

Benjamin-Loison commented 1 month ago

From a legal point of view people naturally navigating on YouTube UI may be more correct, see https://discord.com/channels/933841502155706418/1018941529265553488/1288909547028615259.

Benjamin-Loison / YouTube-operational-API

Monetize this tool #226