Mitch Downey :pci:

open source developer @ podverse.fm / fediverse / foss / javascript / node.js / podcasts / react / react native / typescript

Public Key

npub1yfr5d05qj3qzz84kntze3c8accyz0unwczgm0fxuypuqdnenkwuqew6g6v

NIP-05 Address

mitch@podcastindex-social.mostr.pub

Profile Code

nprofile1qqszy36xh6qfgsppr6mf43vcur7uvzp87fhvpydh5nwzq7qxeuem8wqpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0ds5pyvdc

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1yfr5d05qj3qzz84kntze3c8accyz0unwczgm0fxuypuqdnenkwuqew6g6v

Show more details

Last Notes

2026-03-19 05:33:00 UTC

- reply

@nprofile…zv63 @nprofile…zpdn @nprofile…cn7u this is also something we could add to Podverse Alpha fairly soon. The main blocker for me would just be a working livestream to dev and test against. Is there a 24/7 feed using this in the wild? Maybe too expensive for someone to run, but I imagine the use case could apply for displaying "Now Playing" info for 24/7 RSS radio/tv stations?
Also, the use case seems like it may be broader than "value" specifically. "Live Info Updates" or something.

2026-03-19 05:28:59 UTC

- reply

@nprofile…zv63 @nprofile…zpdn @nprofile…cn7u yes, I would like to support it.
re: a schema, a small suggestion I have would be, where possible, to follow patterns that align with RSS naming conventions?
channelTitle
itemTitle
Not sure what else this could apply to. Not strictly required. Just an idea.

2026-03-04 18:38:41 UTC

- reply

@nprofile…8zna interesting, I assumed this flip happened years ago. I barely know anyone who listens to AM radio anymore.

2026-02-21 17:55:44 UTC

- reply

@nprofile…zpdn I basically burned out on open source development from 2024 to early 2025...feeling much better now tho, and LLM advancements in the meantime have made things even more exciting and easier.
ActivityPub is still in the picture. I think there hasn't been much discussion lately around AP as a solution for boost metadata? @nprofile…lm4c could speak to AP much better than me. I'm still getting caught up.
I do think, broadly, LLM is making it easier to adopt robust standards like AP.

2026-02-07 00:36:33 UTC

@nprofile…8zna I was asking LLM some questions about value tag implementations, and look where it decided to go for source documentation!
https://cdn.masto.host/podcastindexsocial/media_attachments/files/116/026/542/933/631/282/original/216cbc49809353d7.png

2026-01-29 20:23:48 UTC

Podverse Alpha update 🛠️
We are nearly ready to implement V4V in Podverse Alpha web...and I am completely out-of-date in terms of what the current practices are.
Can anyone direct me to resources I should follow for a web-based Bitcoin V4V implementation? Is the standard still keysend? Are there multiple types of implementations we should support by now?
#podverse #pv2

2026-01-26 16:30:37 UTC

- reply

@nprofile…x5pc ok, I think I get it now. The Mostr bridge is silently mapping that npub it assigned to me to mitch@podcastindex.social in the API requests it makes with the fediverse. I imagine if we could see the Mostr API requests we would see our original handles in there along with the npub it makes visible in the message text.

2026-01-26 15:31:14 UTC

- reply

@nprofile…x5pc looking at hellthread website, I don’t think my Fediverse account even has a nostr npub id? Everyone else on the thread does though. Don’t understand how or where I’m tagged on this stuff.

2026-01-26 15:28:44 UTC

- reply

@nprofile…x5pc what’s weird is I don’t even see myself tagged in the threads I’m getting notified for. They haven’t happened again since last night so ig I’ll just ignore it for now.

2026-01-26 02:18:25 UTC

I'm getting notifications containing weird spam that makes no sense on here. Anyone else having that? How do you stop it?

2026-01-25 20:20:18 UTC

- reply

@nprofile…peg3 ah yes, well you are right that npm seems to have something similar to Rust in this respect.

2026-01-25 20:18:05 UTC

- reply

@nprofile…peg3 the pv packages it consumes (and whole infrastructure) will be bumped to 5.0.2, even if they don't have changes.
A major trade-off there is inefficiency, but from what I understand, we could add more robust tooling for monorepo handling later.
NPM has a feature called "workspaces" which appears to drastically simplify our module linking and dev process for a monorepo setup, vs our 10+ repo setup previously.
Anyway...it's a grand experiment lol but I think it's looking good so far

2026-01-25 20:15:20 UTC

- reply

@nprofile…peg3 thanks for the advice. After more research however, I've decided to go forward with the monorepo. I think the trade-offs of improved LLM context handling, easy of "quickstart" development, and simplification of open source contribution and CI are going to be worth it for our very large infrastructure.
One rule that is simplifying monorepo for us is currently we're planning on keeping all the packages and apps on the same version. Ex. if podverse-web updates from 5.0.1 to 5.0.2

2026-01-23 21:30:47 UTC

- reply

but I'm also thinking it may be easier for open source contributors to create PRs for Podverse if they can just create 1 PR per feature. Right now, a typical feature update will require changes to multiple repos. It's a clean process imo, but it does add significant overhead to PR creation and reviews.
Anyone have experience and opinions on monorepos? Worth the switch in our case, or could it create more problems than it is worth?

2026-01-23 21:28:29 UTC

As the Podverse infrastructure has grown to 10+ repos, and given how LLM seems to perform better with all its context in one place, and it would be easier to save llm prompt history in one place...I'm wondering if we should make a major switch to a monorepo 😣
I've never done a monorepo before. If it were just LLM, then we may not need it (we could create yet another repo...podverse-llm or something) (cont)

2026-01-22 18:43:29 UTC

- reply

@nprofile…cyex awesome. makes sense to me. thanks!

2026-01-18 05:11:40 UTC

- reply

@nprofile…tsmm yeah that sounds right. The exploit in that case is limited to only that specific browser though, since the server has to grant the httponly status to that device. Without httponly, the JWT value could be extracted and set in a different client/device and used freely from there.

2026-01-17 16:14:59 UTC

- reply

@nprofile…tsmm no, a web browser can't set an httponly cookie. What the LLM did was modify the existing login endpoint in our router API so that it would return the JWT in a response body, instead of setting it as a cookie.
The browser then received the response body, and set it as a non httponly cookie in the browser.
And yes, httponly is needed on cookies to prevent XSS attack vectors. When httponly is true, browsers prevent the JavaScript runtime environment from being able to read it.

2026-01-17 13:32:11 UTC

- reply

This is an egregious and easily avoidable security vulnerability. JWTs should not be set client-side in a browser or they can be scraped by 3rd parties.
The actual solution was to simply add ‘withCredentials: true’ in the request helper’s options. If we were flying in blind without knowing that, it would have appeared like the LLM gave us a valid solution that worked, but in reality it was a critical mistake.
I think the models will probably get smart enough to avoid this particular error tho

2026-01-17 13:27:23 UTC

I was showing my brother how to do some agentic development, and it went great, except for one notable problem.
We were adding a Login page to a website. The auth endpoints were already written. It generated the form, and made the request to the endpoint, and logged-in…
But the HttpOnly cookie wasn’t set in the browser. So we asked LLM for help debugging.
The solution it wrote was to return the JWT as a token in the browser, then set it in a cookie on the client-side ☠️

2026-01-17 12:11:25 UTC

Guys…I have no idea what people mean when they talk about multi-agent LLM development workflows.
I’ve been using Cursor with “Auto” mode, basically prompting one feature at a time, and I’ve been thrilled with the results.
But I keep seeing ppl talk about multi-agent and I can’t picture what they’re working on. The LLM takes like 10-30 seconds to complete my prompts. Where is there even time to type more prompts before it completes? I can’t tell how much is larp and how much is my own ignorance

2026-01-14 16:45:02 UTC

- reply

@nprofile…05ue it is! Web only though. I would say it’s about 80% finished.
I’m hoping that leaning into LLMs will help me finish the mobile app faster than I expected…we can’t fully release nextgen PV until mobile is finished since we need a major data migration. Trying to make the website and infrastructure a standalone open source project before then tho.
https://alpha.podverse.fm

2026-01-14 14:43:53 UTC

- reply

@nprofile…05ue the new Podverse we are working on has alternate enclosure support built-in from the ground up. We will try to provide the best alternate enclosure functionality we can for your feeds.

2026-01-14 09:24:27 UTC

- reply

I still don’t understand all the common terminology with LLM development I’m seeing thrown around though. Problem could be my own lack of knowledge.

2026-01-14 09:17:50 UTC

I’m finding trying to keep up with LLM advances for coding pretty disorienting…on one hand there are some real and magnificent breakthroughs, and on the other there is a lot of useless slop masquerading as breakthroughs.
Improved LLM models are already accelerating my productivity by several times, but I’m still seeing enough errors and anti-patterns, that I have a hard time believing the “you can do it all with prompts” and “you don’t need to review every line” camp isn’t being naive.

2026-01-12 05:45:53 UTC

- reply

@nprofile…k8p7 awesome. Thank you!

2026-01-10 23:53:22 UTC

- reply

@nprofile…peg3 do you use a full CLI approach? I haven’t made that leap yet, and with how I’m using LLMs at least, it doesn’t seem like they’ve eliminated the need for manual edits in all cases.

2026-01-10 06:23:53 UTC

- reply

@nprofile…nh7l yeah, and it's important to be cautious. I still reviewed all the code, and there were some issues with it, and some choices it made that were not my preference...but even after the manual changes I needed to make, it saved me so much time, it's crazy.

2026-01-10 04:41:10 UTC

- reply

The cost of using Opus 4.5 is substantial. A long 12 hour day of work cost me almost $50 in Anthropic API credits.
Since we haven't turned a profit yet with Podverse, that comes out of my pocket and is not easy to sustain, but the time and energy savings make it worth it.
I'm currently trying to replace Opus 4.5 in my setup with GLM 4.7, with is the highest rated open source / MIT LLM for coding available today. API credits are 10% the cost of Anthropic's, so hopefully it will be good enough.

2026-01-10 04:36:51 UTC

I don't know if there is anyone left who is skeptical...but the LLM updates over the past couple months for programming are a giant leap forward, and it seems like we've crossed an inflection point for the industry.
Using Opus 4.5, I was able to get what would have been 5 days of work for me finished in 1 day.
One stage of my work included writing 8 paragraphs of instructions for Opus, and it wrote hundreds of lines of code across 7 repos, and was ~99% accurate and finished after a few minutes

2026-01-10 04:28:49 UTC

- reply

@nprofile…k8p7 I don't see an artist RSS feed for Hartlight in Podcast Index:
https://podcastindex.org/search?q=hartlight&type=all
Our search directly queries PI API, so if a feed is not discoverable there, it won't be discoverable in PV.
I see that the album feed has a publisher tag that is pointing to the artist feed, which is great, but we have as a requirement that only feed that are in PI can be added to our database.
Does Fountain typically add artist feeds to PI? Or does PI not seek to include them?

2026-01-09 09:00:35 UTC

- reply

@nprofile…td70 @nprofile…7td6 @nprofile…0fv8 @nprofile…peg3 we don’t have an advanced log implementation at all right now. It’s just STDOUT to files from docker containers onto the host machine. I run all the log statements through Winston helpers tho with the intention of doing something more useful with them later. I don’t know what a good logging system looks like yet tbh

2026-01-07 03:35:07 UTC

- reply

@nprofile…peg3 actually...I am just realizing that Unified Push is only compatible with Android. I thought maybe it would work with web also...
Not sure if there would be value in letting people change UP settings through the website. Seems like something that should be configured within the mobile app on each Android device.

2026-01-07 03:11:32 UTC

- reply

@nprofile…peg3 thanks! I'll see how far I can get, and may have some questions for you.

2026-01-07 02:21:05 UTC

- reply

Next, I will try to add Unified Push notification support, so listeners can use their own open-source, self-hosted push notification systems.

2026-01-07 02:20:13 UTC

Podverse Alpha Update 🛠️
Push notifications have been added to Podverse Alpha web.
Listeners can configure 3 different types of notifications:
- New item (episode or music track)
- Livestream scheduled
- Livestream started
They can also set different settings for different feeds.
This was a lot of work to add. The code includes not only support for web, but an orchestrator to send notifications to also iOS and Android (when the mobile app is ready).
https://alpha.podverse.fm
#podverse #pv2
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/851/379/637/421/016/original/a908a9d29f10b179.png
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/851/381/592/141/112/original/78f15838e0804554.png
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/851/383/779/817/232/original/b4031c4fda251068.png

2026-01-03 20:03:51 UTC

- reply

@nprofile…7td6 @nprofile…8zna thanks! Sounds like it should be a simple fix. We just need to prioritize the artwork field over the image field from PI API?

2026-01-02 22:48:44 UTC

- reply

@nprofile…km5y @nprofile…8zna can you link me to podcast you identified the issue with?

2026-01-02 21:10:47 UTC

- reply

@nprofile…km5y @nprofile…8zna thanks! I’ll look into it.
We have a helper function that should use podcast:image, itunes:image, image in that priority order. I must need to fix that.

2025-12-30 21:40:12 UTC

- reply

@nprofile…5yxp looks good 👍 should work fine

2025-12-30 21:37:48 UTC

- reply

@nprofile…8nsk @nprofile…5yxp @nprofile…y5nd ok, as long as it has to be a UUID, then that won't require a lot of changes on our side.
right now our channel.podcast_guid column in postgres is a UUID data type. i have some extra validation before updates are saved to confirm the UUID is UUIDv5, but i can replace that with a check for any valid UUID type.

2025-12-30 21:04:18 UTC

- reply

@nprofile…8nsk @nprofile…5yxp @nprofile…y5nd hmm I designed our data schema to match the PI namespace docs, and it seems to specifically say UUIDv5 for <podcast:guid>. If other UUID types are valid for <podcast:guid> I need to make substantial changes.
https://github.com/Podcastindex-org/podcast-namespace/blob/main/docs/tags/guid.md

2025-12-29 12:04:17 UTC

- reply

@nprofile…5yxp but I understand that people who are early adopters of a tag like remoteItem may be manually creating these feeds by hand, which makes it possible for something like copying 35 characters of a podcast guid when a podcast guid needs 36 characters in UUIDv5 format to adhere to the P2.0 spec. This is super easy for an RSS generator to programmatically validate and make sure never happens…but not easy when people may be typing or copy/pasting fields like this manually by hand.

2025-12-28 19:44:58 UTC

- reply

@nprofile…0lgv I think I’ve made progress on this. The he.decode line in the parser config was stumping me for a while. Things seem to be working when I remove it. Now I gotta figure out what problems removing that could cause…

2025-12-28 08:40:55 UTC

- reply

@nprofile…0lgv I tried to make progress with an upgrade from v3 to v4, but I have not been able to figure out how the new validated and parsed output format is supposed to work.
I imagine many helper methods that you wrote would need to change. I'm open to doing this work, but I don't have a clear direction on how to proceed yet. If there's any insight you can offer I would appreciate it, but if you're too busy no problem.

2025-12-27 00:17:34 UTC

- reply

@nprofile…dzgg @nprofile…y5nd we should be able to integrate with the Rust parser outputs. I think we may want a layer of Typescript types to go with the output. That’s just a data mapping exercise tho, and not strictly necessary.

2025-12-26 18:38:16 UTC

- reply

@nprofile…dzgg but also...partytime just seems really stable for our purposes, so I'm not sure if it is worth it for us to break from our full-stack JS pattern to incorporate feedparser/Rust. Also partytime uses Typescript types which is good for our builds and workflow.
Anyway, if no one else is eager to maintain partytime, I think I will fork it tomorrow, make the changes needed, and maybe longer term start a migration to feedparser. I see you made some PRs that are not merged and I can take a look.

2025-12-26 18:35:06 UTC

- reply

@nprofile…dzgg I have seen it. I'm open to using it at some point, but podcast-partytime works well with our full-stack JavaScript infrastructure.
I'm thinking it would take me a day of work to add the missing pieces we need to podcast-partytime, but switching to feedparser which is written in Rust would be a bigger initiative.
partytime logic is only used in podverse-parser/src/lib/compat and podverse-parser/src/lib/rss, so swapping it out doesn't seem crazy difficult
https://github.com/podverse/podverse-parser/blob/v5-develop/src/lib/rss/channel/channel.ts

2025-12-18 20:31:50 UTC

- reply

@nprofile…5yxp sorry, too early. Anything on Alpha could be deleted later. There will be a complex migration process from PV1 to PV2 when it finally goes live, and the alpha data won’t be part of that (hence the disclaimer and fact we’re not charging for alpha memberships).

2025-12-18 16:34:21 UTC

- reply

@nprofile…5yxp @nprofile…8nsk reorganizing your queue with thousands of items a huge operation bc every rows position in your queue would need to be updated when 1 row changes 🤦‍♂️

2025-12-18 16:33:10 UTC

- reply

@nprofile…5yxp @nprofile…8nsk it was probably due to an “auto add new episode to queue” feature being turned on, which was possibly buggy.
The new history / now playing / queue system is hopefully improved over old version. In PV1 those were 3 tables. In PV2 it’s one table with a list_position column:
history < 0
now playing = 0
queue > 0
It uses large decimal position numeric data types to easily insert rows where we want instead of an integer type. Old version used an integer type which would make

2025-12-17 18:31:39 UTC

🛠️ Podverse Alpha Update 🛠️
IT'S A BIG ONE
Started to add the Albums, Album, Tracks, and Track pages for Music playback.
The UX also splits the Queue, History, and Playlist into Podcasts and Music.
Any feedback or suggestions for improvement would be appreciated!
https://alpha.podverse.fm
#podverse #pv2
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/736/331/378/281/651/original/2e21b660fcb53be1.mp4

2025-12-17 05:35:14 UTC

- reply

@nprofile…8nsk @nprofile…wsn5 @nprofile…5yxp if tags is a good fit, i'll use them.

2025-12-17 01:13:20 UTC

- reply

@nprofile…wsn5 @nprofile…5yxp
> Wavlake adopted the category tag, implemented wrongly, but whatever, it works.
What did they handle incorrectly? I see this example from The Underclass:
<podcast:category text="Rock">
<podcast:category text="British Invasion"/>
</podcast:category>
is that incorrect?
https://wavlake.com/feed/music/fa16edd9-d3d2-4c6f-a6e1-c9d43307a0bf

2025-12-17 01:04:47 UTC

@nprofile…5yxp @nprofile…wsn5 i've asked this before but forget 🤦‍♂️
is there something comparable to "categories" or "genres" in use for P2.0 music feeds?

2025-12-16 21:18:48 UTC

- reply

@nprofile…wsn5 @nprofile…5x2e @nprofile…5yxp @nprofile…vwpk this is good to know. thanks!

2025-12-15 05:08:51 UTC

- reply

I *think* I'm clear to start adding the Music pages tomorrow.

2025-12-15 05:08:17 UTC

Podverse Alpha update
Added a new database table named on_demand_parser_event, which will keep track of when a listener adds a feed to our directory, and help prevent overuse / spamming of our "add feed to directory" feature.
Not an exciting update from a listener's standpoint, but an important step, since we are giving listeners the ability to trigger RSS parsing on our servers in PV2, and that can be a data and resource heavy operation.
#podverse #pv2

2025-12-13 04:03:18 UTC

- reply

Podverse Alpha Web progress:
- Redesigned the Subscribe button layout.
- Added a "Podcast Settings" page and a "Check Feed for Updates" button to it. That button lets users manually trigger a feed re-parse in our directory.
- Some misc style improvements for mobile devices.
I did not start on the music pages yet, but that's next on the list.
Changes are not deployed yet. I think I will wait until the music pages are finished before deploying next, possibly Sunday night.
#podverse #pv2
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/710/233/783/352/045/original/c94a1e867c9b6374.png

2025-12-11 02:46:54 UTC

- reply

@nprofile…8nsk it seems like a natural fit to me, but the fallback I mentioned doesn't seem bad in the meantime. I'm curious if other people would want it.

2025-12-10 23:33:33 UTC

Does Podcasting 2.0 have a way to link an "ended" livestream, to the recording of that broadcast within an RSS feed?
I imagine it could take the form of a guid in the <podcast:liveitem>, and then when the broadcast is finished, that same guid is included in a tag for the <item> within the same feed.
If we had something like this, then an "ended" livestreams page like this could make clicking on an ended stream redirect to the item (if it exists).
Or is this a behavior we'd rather avoid?
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/697/865/986/817/684/original/e956d32fcd84358a.png

2025-12-10 06:21:28 UTC

- reply

The pagination widget on the bottom showing too many pages is a known issue. Will be fixing that soon.

2025-12-10 06:20:17 UTC

- reply

On a better note, here is a work-in-progress gif preview of the Livestreams page I worked on today for Alpha.
Next I need to add a Livestream page, then we can deploy these changes.
Albums and Tracks pages should be finished shortly after. They will basically be copies of the existing Podcasts and Episodes pages, so I'm thinking it'll only take a day.
#podverse #pv2
https://cdn.masto.host/podcastindexsocial/media_attachments/files/115/693/820/518/179/316/original/2ad33a1bc32bf7d6.mp4

2025-12-10 06:16:12 UTC

⚠️ Podverse Alpha Update ⚠️
I'm shutting off the Alpha website for the night, as there is a critical security vulnerability with the version of Next.js it is using (CVE-2025-55182).
Should have it redeployed and patched tomorrow.
Good video explaining the CVE: https://www.youtube.com/watch?v=s81dVUM-cQM
Thanks to @nprofile…ezja or notifying me.
#podverse #pv2

2025-12-09 04:19:56 UTC

- reply

@nprofile…t46n thanks for the heads up! I change the rate limit rules for function.

2025-12-06 19:13:10 UTC

Decided to merge "audio" and "video" into one grouping. Rewriting the API and modules to be able to handle it.
I think the primary groupings listeners will see will be Podcasts and Music.
#podverse

2025-12-05 14:33:37 UTC

- reply

@nprofile…5yxp it helps! I'm leaning towards merging the experiences in the UI...

2025-12-04 19:16:12 UTC

- reply

@nprofile…4t4s I’ll be working on the pagination funkiness. Relatively light fixes to make.

2025-12-04 19:15:15 UTC

- reply

@nprofile…4t4s maybe we should make subscribe a bit smaller, and fit it along side the other buttons on the bottom left of the header?

2025-12-04 19:11:22 UTC

- reply

@nprofile…4t4s if you search for your podcast again, it should direct you to the podcast page this time.

2025-12-01 20:08:01 UTC

Update on Podverse 2.0...first version of the Alpha website should be publicly accessible this week.
Technically could be shared now...but ironing out some things and finishing some pages before that.

2025-11-29 05:16:18 UTC

- reply

@nprofile…nhmd thanks! I’ve bookmarked this and will look into it when we’re ready to write the mobile app.

2025-11-19 23:05:54 UTC

- reply

@nprofile…nhmd can you remind me what NWC is? 😅

2025-11-16 19:32:22 UTC

- reply

@nprofile…5yxp yessir! Music will be supported as a primary feature, instead of how it is in PV1, where it's kind of an afterthought shoe-horned in.

2025-11-12 00:38:51 UTC

Does anyone know of a 24/7 podcasting 2.0 compatible video livestream, I could use for dev purposes?
Adding video livestream support to Podverse 2.0. I'll have to learn how to setup my own livestreams on localhost sooner than later, but wondering if there's already a video live feed people know of.

2025-04-17 23:33:52 UTC

- reply

@nprofile…2d4u @nprofile…ye27 @nprofile…3ram @nprofile…7xwn @nprofile…7te3 @nprofile…r3p6
I wish I had a good answer, but I don't remember how lnaddress works right now, and I'm guessing Podverse doesn't support it.
I'd expect to support whatever value tag options are available in PV2, and I assume lnaddress will be a part of that, but as for current adoption, I'm afraid PV does not support it afaik.

2025-03-21 19:18:17 UTC

- reply

@nprofile…3ak2 yooo, love the music! this is the type of chill music i listen to while coding.

2025-03-19 17:29:43 UTC

- reply

7) After 30 days, track event data is automatically deleted from our system. (The aggregate total stats persist, but there will be no information that ties the data back to specific Premium member accounts.)
8) As always, we will never share private user information with 3rd parties. This "tracking" system exists only for popularity metrics. Podverse will never add advertisements, so there's no reason for us to collect anything more than the bare minimum information.

2025-03-19 17:29:36 UTC

- reply

4) Premium members can opt out of having their listens counted in popularity stats if they want.
5) Popularity statistics will update more frequently in PV2 compared to PV1, so the stats you see will be closer to real-time data.
6) Tracking data collected is minimal. Unlike in PV1 (which uses a self-hosted service called Matomo), no device info will be collected. The data we'll collect is simply "did Premium member listen to this or not".

2025-03-19 17:29:15 UTC

🛠️ Podverse 2.0 Update 🛠️
@nprofile…nrc9 and I have spent the last week working on a new stats / popularity sorting system for PV2. Here is some info about the new popularity system:
1) Listeners can sort by most popular daily, weekly, monthly, and all-time statistics.
2) Listeners can sort podcasts, episodes, clips, playlists, and users by popularity.
3) Only tracking events by paid Premium members will be counted in the stats. This is intended to prevent spam and manipulation.

2025-03-15 03:14:46 UTC

- reply

@nprofile…h82d sorry, another question...
Podverse only needs to know which RSS feeds have value tags in PW, aka which feeds do NOT have value tags in the RSS.
I'm aware of this PI endpoint: /podcasts/bytag?podcast-value=true&max=5000
However, won't this return ALL feeds that have value tags, and not just the feeds that have value tags external to the feed in PW wallet?
Should PV just always use PI API as the source of truth for value tags, and not try to use value tags we parse in RSS feeds?

2025-03-15 02:26:10 UTC

- reply

@nprofile…h82d I see there is a byepisodeguid endpoint, but it looks like it is for finding multiple value tags for episodes from ONE podcast at a time, but for Boostagram Ball, the remoteItems are each technically a different podcast I believe.
https://podcastindex-org.github.io/docs-api/#post-/value/batch/byepisodeguid
If iterating is the way to go, I think that is fine, I just want to make sure we don't get rate limited.
I could also write a proxy endpoint with a PI API key, and make these requests through our API, but not sure if that's recommended.

2025-03-15 02:24:08 UTC

- reply

@nprofile…h82d in PV2, I am thinking of having the client-side apps make requests for PW hosted value tags as needed. My questions are:
1) Is there any rate limiting I should be aware of?
2) In the case of a feed like Boostagram Ball, with many remoteItem value tags within it, should I just iterate over the remoteItems, making requests to PI API for each one? Or is there a batch endpoint for cases like this?
(cont)

2025-03-15 02:20:46 UTC

@nprofile…h82d I have a question about Podcaster Wallet API practices.
In PV1, during our server-side parsing, we make requests to PI API for the value tags for a feed that are in PW. This has resulted in occasional issues, as sometimes our db has an outdated reference for PW hosted value tags (especially since I believe when PW value tags update, it does not count as a "feed update" for our servers can be notified of).
In PV2...

2025-03-13 05:55:08 UTC

- reply

thank you @nprofile…6746 ❤️ honestly we should be paying you for all the help you've given us

2025-02-22 01:40:46 UTC

- reply

@nprofile…h82d @nprofile…nrc9 finally we're at the ItemAbout service:
https://github.com/podverse/podverse-orm/blob/v5-develop/src/services/item/itemAbout.ts
and that uses an abstracted class BaseOneService (unclear name...it means one item can only have one itemAbout row):
https://github.com/podverse/podverse-orm/blob/v5-develop/src/services/base/baseOneService.ts#L25
the transaction handling logic happens in that reusable base class.
a lot of info and very javascripty 😬

2025-02-22 01:37:40 UTC

- reply

@nprofile…h82d @nprofile…nrc9 I can't remember any raw sql I've written for pv2 yet, except for the table / index / etc creation logic.
the outermost parser function looks like this:
https://github.com/podverse/podverse-parser/blob/v5-develop/src/lib/rss/parser.ts#L43
going deeper, item parsing looks like this. the transactionalEntityManager is what makes the orm queries within these functions happen in a transaction:
https://github.com/podverse/podverse-parser/blob/v5-develop/src/lib/rss/item/item.ts#L197
and here's what the ItemAbout parsing looks like:
https://github.com/podverse/podverse-parser/blob/v5-develop/src/lib/rss/item/itemAbout.ts
and...

2025-02-21 23:32:59 UTC

- reply

@nprofile…pvn3 sorry, i misremembered. it's not exactly a typescript issue. if i don't use static then i can't use the Controller classes in the way shown in the first screenshot.
i suppose i could handle it by instantiating the class first, like in the second screenshot.
i'm not sure the change will be consequential, but maybe it's a better practice than using static.
https://cdn.masto.host/podcastindexsocial/media_attachments/files/114/044/474/793/038/965/original/10bd497799d2baf2.png
https://cdn.masto.host/podcastindexsocial/media_attachments/files/114/044/480/389/759/554/original/0cdd7e7d7fac2f6e.png

2025-02-21 23:15:44 UTC

- reply

my usage of static methods may be a little sus here. i was having trouble getting typescript to work without static. i'm not sure static is a problem here though

2025-02-21 23:13:21 UTC

This is what a lot of my programming looks like today:
"write a QueueResourceItemController with methods that correspond with each of them in the QueueResourceItemService"
and from that Copilot spit out 102 lines of code. In this case it happened to work first try, using patterns already set in other files for the project.
You still have to review the code manually, but its amazing how well it handles some types of work. It's kinda like using a calculator for math instead of pen and paper.
https://cdn.masto.host/podcastindexsocial/media_attachments/files/114/044/367/752/029/103/original/f3e97a874c63f9f3.png

2025-02-21 20:58:59 UTC

- reply

@nprofile…h82d @nprofile…nrc9 there are a lot of repos that comprise this infrastructure
podverse-api
podverse-external-services
podverse-helpers
podverse-ops
podverse-orm
podverse-parser
podverse-queue
podverse-workers
Each of those has a v5-develop branch with the current "Podverse 2.0" code.
Future repos we expect to have PV2 code:
podverse-ansible
podverse-rn (mobile / react native)
podverse-web
It's big freakin infrastructure 😵‍💫 the separation of concerns help keep things manageable and portable though

2025-02-21 20:54:21 UTC

- reply

@nprofile…h82d @nprofile…nrc9 yes the sql file is. Here is the combined "initialize database" file.
https://github.com/podverse/podverse-ops/blob/v5-develop/database/combined/init_database.sql
The separate files that get combined into that big one are here:
https://github.com/podverse/podverse-ops/tree/v5-develop/database/migrations

2025-02-21 17:36:59 UTC

- reply

@nprofile…g5pf @nprofile…6746 @nprofile…7xwn @nprofile…r3p6 @nprofile…ye27 @nprofile…2802 @nprofile…7te3 @nprofile…2d4u
It's definitely a clickbait title. I'm willing to hear them out though. As long as people aren't deliberately misrepresenting things I'll listen to the skeptics.

2025-02-21 17:31:45 UTC

- reply

@nprofile…6746 @nprofile…7xwn @nprofile…r3p6 @nprofile…ye27 @nprofile…2802 @nprofile…7te3 @nprofile…2d4u
Thank you brother. I don't take "PC 2.0 is dead" seriously lol all it is is a clickbait title. Some of the 2.0 tags already have mass adoption, and an adopted standard can't be "dead."
I'm fine with people saying it though because they usually mean it to make other points about the future of PC 2.0. They'll see soon enough it's alive and growing.

2025-02-21 17:04:18 UTC

- reply

5) The ability to make medium-specific playlists. For example, you could make a playlist that any type of item can be added to, or one that is just for podcasts, or just for music, or videos, etc.
Next on my list is the improved Queue /Now Playing / History handling. It will have a lot of similarities to the new Playlist feature.

2025-02-21 17:00:18 UTC

PV2.0 update:
The Podverse API work is ~80% finished.
This week I've spent most of my time writing the new Playlists feature.
Here are some improvements you can expect PV2.0:
1) The ability to add episodes, chapters, clips to playlists.
2) The ability to add items from your custom "add by RSS" feeds to playlists.
3) Multiple privacy settings. Playlists can be public, only with link, or private.
4) Improvements to scalability which should allow playlists to support thousands+ items.

2025-02-21 16:53:17 UTC

- reply

@nprofile…6746 @nprofile…7xwn @nprofile…r3p6 @nprofile…ye27 @nprofile…2802 @nprofile…7te3 @nprofile…2d4u
thanks Chad! you've helped our apps tremendously too. you're basically an advisor to all our teams.

2025-02-15 22:59:16 UTC

- reply

@nprofile…pk3p I think so? I've made like 4 Electron apps but haven't used it in a few years.
Do you have any ideas for what the benefits of an Electron app would be? The only one I can think of off the top of my head is we could make it easier to download episodes from desktop.
Another benefit would be we can get around CORS restrictions so we can parse RSS feeds in browser, but we're planning a server-side solution for that so CORS shouldn't be a problem in PV2.0.

2025-02-14 23:49:56 UTC

PV2.0 status update:
The goal of this infrastructure is to support ALL podcasting 2.0 features, and full feature parity between web and mobile.
Work on the new database and parser system is mostly finished.
Progress on the data API is going well. Added basic channel and item (aka podcast and episode) endpoints, and a new auth system for login.
The improved separation of concerns of Podverse modules (10 REPOS 😳) has made coding PV2.0 much cleaner and easier.
https://github.com/podverse/

2024-10-25 19:23:25 UTC

- reply

@npub1lh5…n08y @npub12x8…mjy3 all good homie lol I just don't want a rumor to go around NA social that Podverse has changed its content policies. That's definitely not true.