no! Now I get "The provided API key is invalid."

not sure how I entered that anymore. How can I reset it or find out what it is?

 

19 minutes ago, tvmainia said:

no! Now I get "The provided API key is invalid."

 

18 minutes ago, tvmainia said:

not sure how I entered that anymore. How can I reset it or find out what it is?

 

The API key is the one you set in the Meiliseach config and should have copied to the Hoarder config.

i'll generate another one, I think i found a link on meilisearch website. I'll report back.

 

Thank you!

@Collectathon I regenerated they key and search is working! Thank you so much again.
Next, figuring out how to set up AI.

Thanks for putting this together. To confirm, is Hoarder-Workers no longer a requirement? I don't see that listed in the Apps. Was about to start the process of setting this up, but wanted to ensure I understood what was needed before diving in.

what is the "NextAuth" and associated secret for?  

8 minutes ago, jpresson said:

what is the "NextAuth" and associated secret for?  

Hoarder will work without setting the NextAuth address. It's used for OAuth, as you would set this to your local or public Hoarder address. The only thing you'll notice by not setting this if using local authentication is that when you sign out, you won't be redirected properly.

Due to trademark reasons, Hoarder has been rebranded to Karakeep. The good news is that both the old and new image repositories will continue receiving updates, so you don't have to update your template immediately.

However, if you'd like to switch to the new naming scheme, you can update your template by changing the repository line:

 

From:

ghcr.io/hoarder-app/hoarder:release

 

To:

ghcr.io/karakeep-app/karakeep:release
 

You do not have to change any other instance of the word Hoarder in the config, but you can if you would like to. Alternatively, you can stop your existing Hoarder container, use the new Karakeep template in CA and move your existing configuration to the new template.

 

Edited April 6, 20251 yr by Collectathon

Hoarder/Karakeep was working like a charme for me, now since about a week ago, links added don't get fetched anymore. My Karakeep dashboard is full of white/emtpy previews with the pasted URL as title - performing a 'refresh' on every item does not get the content either. When restarting the docker, Karakeep sometimes refreshes the empty content itself, but it has worked for many months and i'm not understanding what's happening here. I'm on 0.24.1 btw and never skipped any update.

7 hours ago, Gekko said:

Hoarder/Karakeep was working like a charme for me, now since about a week ago, links added don't get fetched anymore. My Karakeep dashboard is full of white/emtpy previews with the pasted URL as title - performing a 'refresh' on every item does not get the content either. When restarting the docker, Karakeep sometimes refreshes the empty content itself, but it has worked for many months and i'm not understanding what's happening here. I'm on 0.24.1 btw and never skipped any update.

Can you please share logs from Karakeep container?

Is there any way to enable a log that it's a bit more descriptive?

I'm setting Karakeep for my family, since Pocket is closing down, and I'm having a hell of a time getting browserless to work.

2025-05-24T10:48:34.992Z info: [Crawler] Connecting to existing browser websocket address: ws://browserless:11312?token=ORckYQ3AW06NpAw

2025-05-24T10:48:34.994Z error: [Crawler][2016] Crawling job failed: [object Object]

I set browserless-v2 and changed the name (removed -v2) and the port to an arbitrary one since 3000 was already used and token matches, so I don't know why it's not crawling correctly.

Edit: Changing it to IP:PORT instead of the name of the container made it start working. Why wouldn't the container name work? We'll never know... I suspect my docker is cursed. 🥲

Edited May 24, 20251 yr by Nirvash

1 hour ago, Nirvash said:

Edit: Changing it to IP:PORT instead of the name of the container made it start working. Why wouldn't the container name work? We'll never know... I suspect my docker is cursed. 🥲

I'm not sure why that is the case. I'm using the docker name as well in a custom network and it works as expected.

I'd understand them not working if they weren't in the same network, but they are... I've had issues with docker networks before so I'll blame it on that.

On 5/10/2025 at 1:29 PM, Gekko said:

Hoarder/Karakeep was working like a charme for me, now since about a week ago, links added don't get fetched anymore. My Karakeep dashboard is full of white/emtpy previews with the pasted URL as title - performing a 'refresh' on every item does not get the content either. When restarting the docker, Karakeep sometimes refreshes the empty content itself, but it has worked for many months and i'm not understanding what's happening here. I'm on 0.24.1 btw and never skipped any update.

This happened to me too. In my case it was because my docker network IPs had randomly changed! No idea why, but essentially the Karakeep container was no longer pointing at the Browserless one as the IP address was no wrong. I don't think I had changed anything on my end, so assume it was a result of one update or another.

Karakeep worked well for some time but now I am getting the following error in the Log whenever I try to do any operation beyond login in which works for existing user.

SqliteError: unable to open database file

at w.values (/app/apps/web/.next/server/chunks/8195.js:67:50497)

... 4 lines matching cause stack trace ...

at process.processTicksAndRejections (node:internal/process/task_queues:105:5) {

cause: SqliteError: unable to open database file

at w.values (/app/apps/web/.next/server/chunks/8195.js:67:50497)

at w.all (/app/apps/web/.next/server/chunks/8195.js:67:50086)

at QueryPromise.all (/app/apps/web/.next/server/chunks/8195.js:67:106941)

at QueryPromise.execute (/app/apps/web/.next/server/chunks/8195.js:67:107064)

at QueryPromise.then (/app/apps/web/.next/server/chunks/8195.js:67:79304)

at process.processTicksAndRejections (node:internal/process/task_queues:105:5) {

code: 'SQLITE_CANTOPEN'

},

code: 'INTERNAL_SERVER_ERROR',

name: 'TRPCError'

}

How can I solve this? Also is there a way to export/backup existing entries and re-install completely?

14 hours ago, chortya said:

Karakeep worked well for some time but now I am getting the following error in the Log whenever I try to do any operation beyond login in which works for existing user.

How can I solve this? Also is there a way to export/backup existing entries and re-install completely?

Could you please share your Unraid template?

Karakeep's database is in the data dir under called db.db. Optimally, you should back up the entire data dir as it contains all the assets (thumbnails, archives, etc). By default, the template sets this location to /mnt/user/appdata/karakeep. You can make a copy of this, remove the container, recreate the container, then add the database back.

On 5/26/2025 at 2:16 PM, danclark said:

This happened to me too. In my case it was because my docker network IPs had randomly changed! No idea why, but essentially the Karakeep container was no longer pointing at the Browserless one as the IP address was no wrong. I don't think I had changed anything on my end, so assume it was a result of one update or another.

Hi,

could you please explain to a newbie how you got it working again?

8 hours ago, linuxdevil said:

Hi,

could you please explain to a newbie how you got it working again?

I have each container on a custom docker network so in the range 172.18.0.x but I hadn't bothered assigning each container a static IP they were set to dynamic. So essentially what had happened for some reason was that some of the IPs on my customs docker network were reallocated. So essentially the Browserless IP was no longer 172.18.0.17 and had changed to 172.18.0.21 so basically Karakeep was no longer looking for it in the right place. I updated the Browserless path in the Karakeep container settings to point to the new IP and it started working again.

Hope this helps.

Edited June 25, 20251 yr by danclark

1 hour ago, danclark said:

I have each container on a custom docker network so in the range 172.18.0.x but I hadn't bothered assigning each container a static IP they were set to dynamic. So essentially what had happened for some reason was that some of the IPs on my customs docker network were reallocated. So essentially the Browserless IP was no longer 172.18.0.17 and had changed to 172.18.0.21 so basically Karakeep was no longer looking for it in the right place. I updated the Browserless path in the Karakeep container settings to point to the new IP and it started working again.

Hope this helps.

That did the job, thanks a lot.

I'm getting the same error as above: error: [Crawler][141] Crawling job failed: TimeoutError: browserType.connect: Timeout 5000ms exceeded.

I have ip:port and container:port still no luck.

15 minutes ago, Rewrap said:

I'm getting the same error as above: error: [Crawler][141] Crawling job failed: TimeoutError: browserType.connect: Timeout 5000ms exceeded.

I have ip:port and container:port still no luck.

I assume 192.168.0.18 is your Unraid IP. I can see that they are both on the same custom network which I assume is a custom bridge network.

With that info, you are currently trying to connect to Browserless via 192.168.0.18:3000 when the port for Browserless has actually been changed to :3012

Either of these addresses should work to get Karakeep connected to Browserless:

192.168.0.18:3012

browserless:3000

Edited July 20, 2025Jul 20 by Collectathon

1 hour ago, Rewrap said:

I'm getting the same error as above: error: [Crawler][141] Crawling job failed: TimeoutError: browserType.connect: Timeout 5000ms exceeded.

I have ip:port and container:port still no luck.

I am having this issue as well. I've tried re-creating Browserless from the default template, as well as adding port 3012 to it (is that websocket, or http?). I do see activity in the Browserless logs when I try to refresh a link in Karakeep, but I see the same error as @Rewrap . I believe this started with the latest Karakeep update. Rolling back to 0.25.0 breaks my Karakeep instance, so I am unable to verify.

UPDATE

I was able to downgrade to 0.25.0 (thank you snapshots), and crawling works as expected now.

Edited July 20, 2025Jul 20 by anibis2

To get it to work for Karakeep 0.26 change the BROWSER_WEBSOCKET_URL to ws://[IP]:[PORT]/chromium/playwright?token=[TOKEN]

i.e. insert "/playwright"

See https://github.com/karakeep-app/karakeep/issues/1758

Edited July 22, 2025Jul 22 by traynier
clarification of karakeep version

9 hours ago, traynier said:

To get it to work for Karakeep 0.26 change the BROWSER_WEBSOCKET_URL to ws://[IP]:[PORT]/chromium/playwright?token=[TOKEN]

i.e. insert "/playwright"

See https://github.com/karakeep-app/karakeep/issues/1758

Are you using Browserless or something else for your scraping? I tried making the above change and I get the same error. I changed it from ws://[IP]:[PORT]/?token=[TOKEN] to ws://[IP]:[PORT]/chromium/playwright?token=[TOKEN]. When that didn't work, I deleted the chromium part and got a playwright version mismatch error.

UPDATE

I saw that the Browserless container I was using from the Unraid appstore hadn't been updated in over a year, so I swapped to the repo that Browserless v2 uses. All is well now with the updated websocket URL.

Edited July 22, 2025Jul 22 by anibis2