Hi, after a long hiatus (life got busy) I've finally had something new to share from tinkering with open-source projects. Hope it could be useful for the community. PS: I hope it's still okay that...
Hi, after a long hiatus (life got busy) I've finally had something new to share from tinkering with open-source projects.
Hope it could be useful for the community.
PS: I hope it's still okay that I'm posting my own blog here. I checked and I think I'm good. But if I'm mistaken, please do let me know if I should stop. :)
I was curious about your set up. I'm in the process of putting together a server for my home as well. Do you use a NAS and point your paperless instance towards that, or do you use storage on the...
I was curious about your set up. I'm in the process of putting together a server for my home as well. Do you use a NAS and point your paperless instance towards that, or do you use storage on the server itself, or something else? I've been a bit paralyzed by choice regarding this, and think I'm leaning towards a "all in one" setup using one machine. Maybe building a NAS later would be worth it for me, maybe not!
For now I use a VPS service for both storage and hosting. I do want to get around to hosting things locally at some point although I am in the same boat as you with the analysis paralysis. Maybe...
For now I use a VPS service for both storage and hosting. I do want to get around to hosting things locally at some point although I am in the same boat as you with the analysis paralysis.
Maybe someone knowledgeable ends up finding this call for help ahah.
Thank you! I am still waffling between an all in one solution and a NAS + Server setup. I have the impression that separating out the functions is a good move but since my needs are so simple an...
Thank you! I am still waffling between an all in one solution and a NAS + Server setup. I have the impression that separating out the functions is a good move but since my needs are so simple an all in one is sufficient
FWIW I'll put in a vote for Unraid. I set it up a couple years ago just on an old desktop and it's slowly been growing ever since. The big selling point to me was it doesn't need all same size...
FWIW I'll put in a vote for Unraid. I set it up a couple years ago just on an old desktop and it's slowly been growing ever since. The big selling point to me was it doesn't need all same size hard drives. I've even moved it into a dedicated server case and upgraded the hardware and all moving the 'server' took was plugging in the same thumb drive. It's been great having the option to just throw up a docker and try something out. (I have paperless-ngx setup too and it's very nice, though I'm pretty lightweight with it.)
TrueNAS and alternatives sound neat but they're a bigger initial lift.
Not sure what you're looking to use it for, but unless you're expecting a several people to use it simultaneously it's not necessary to have dedicated machines for every piece. Useful for fun and redundancy though of course. I have a Pi5 as a second basket for things that are annoying to go down, pihole/syncthing/home assistant basically.
This sounds great. I already have an RPi4 running Adguard and synching. Never got into HA, but have dreamt about it. I think setting up a dedicated unRAID machine would be very nice. Just making...
This sounds great. I already have an RPi4 running Adguard and synching. Never got into HA, but have dreamt about it. I think setting up a dedicated unRAID machine would be very nice. Just making sure I understand: do you have the unRAID machine working as just a NAS, or also hosting your services? I basically just want a few simple things running. It would only be me, and maybe 1 other person ever using it, so multiple machines is almost certainly unnecessary.
Edit: Services I would want are paperless, immich, and jellyfin, with other minor things
It's operating as both a NAS and general host. I could split the two up but it's really nice to be able to have effectively unlimited storage for whatever little dockers I want to play with since...
It's operating as both a NAS and general host. I could split the two up but it's really nice to be able to have effectively unlimited storage for whatever little dockers I want to play with since I have TB to spare.
I have a long list at this point but it's grown as I try something out for a bit and get it stable and then hear about something and add more.
I've got Plex as my main media sharing platform because it's easier to setup for family but I've spun up Jellyfin at the same time to test it out. Immich has been working great for the last few months, though I haven't pulled the trigger on deactivating Google Photos, mostly out of laziness. Also have calibre-web-automated and audiobookshelf for their respective formats. Nextcloud is nice for a self-hosted cloud replacement. Traccar is one I've been messing with because I really like using Google Maps to see where I've been. So far it's fine but I need to tweak it more and see about importing historical data. FreshRSS is nice to keep up with webcomics in a more organized manner. SearXNG is the engine the Kagi search engine uses, but for free. Heimdall is a nice homepage where I can put 'bookmarks' so I don't have to remember port numbers for all my services. There are a few game servers idling in the background, corekeeper/minecraft/etc. I also have ArchiveTeam-Warrior and Boinc dockers going to give me the warm and fuzzies.
Are you averse to using cloud storage? That's my plan, but wondering if I'm not considering all the things. I figure with encryption using my own key, I don't really care if AWS has my data, and...
Are you averse to using cloud storage? That's my plan, but wondering if I'm not considering all the things. I figure with encryption using my own key, I don't really care if AWS has my data, and would avoid the hassle of having to manage my own hardware and backups.
I think I am averse to cloud data, certainly more than most at baseline. I only want to upload my photos and documents to a service after they have been encrypted with something like cryptomator,...
I think I am averse to cloud data, certainly more than most at baseline. I only want to upload my photos and documents to a service after they have been encrypted with something like cryptomator, Borg, etc. it is a more extreme position, so I understand it's not for everyone. I think if it's not on my own hardware, I don't own it!
I went with cloud storage just to get started. It is a pain to encrypt everything so I one day (hopefully within the year) wish to move to my own on-prem server for my apps.
I went with cloud storage just to get started. It is a pain to encrypt everything so I one day (hopefully within the year) wish to move to my own on-prem server for my apps.
Maybe the what option is on premise live storage with encrypted backups sent to cloud for redundamcy. Cold storage is pretty cheap and should your house ever burn down...
Maybe the what option is on premise live storage with encrypted backups sent to cloud for redundamcy. Cold storage is pretty cheap and should your house ever burn down...
FWIW, I have a Synology NAS and I love it. It mostly self-manages, it runs Docker containers, and you can set them to be accessible over the web. I bet it would be great for something like this.
FWIW, I have a Synology NAS and I love it. It mostly self-manages, it runs Docker containers, and you can set them to be accessible over the web. I bet it would be great for something like this.
I've thought about going with an off-the-shelf solution but am actually excited to build something myself. It's been a while since I've built a computer and I used to enjoy it a good bit. I have a...
I've thought about going with an off-the-shelf solution but am actually excited to build something myself. It's been a while since I've built a computer and I used to enjoy it a good bit. I have a few drives hanging around that I could use in a selfbuild that might not be Synology compatible. Thank you for the recommendation though, these solutions have gotten very good!
Great article! Paperless-ngx is so cool, I've been using it for a little while now myself but only for digital documents for now. But I recently migrated my setup to a local system and with that I...
Great article!
Paperless-ngx is so cool, I've been using it for a little while now myself but only for digital documents for now. But I recently migrated my setup to a local system and with that I explored the email processing feature a little more and I came up with a pretty nice little automated workflow.
The first thing I did was I created a dedicated email address for Paperless, just so I don't have to worry about anything getting imported that shouldn't. And then I made the Mail Rule in Paperless check for only emails coming from my personal email address with PDF attachments. This is pretty standard I think for automated mail importing.
But then in my mail client of choice, Thunderbird, I added this message filter. So when I archive an email and it has an attachment, and the sender is in my "Paperless Correspondents" address book in Thunderbird, it'll forward that email to the Paperless address and get automatically imported. Isn't that so cool!! I found it so neat when I figured out Thunderbird can do all this in an easy way.
I also run my setup with the consume directory available via Samba so should I want to bulk import something it's easy enough to do it there. Though the web UI works too in limited numbers.
Yeah to be fair I don't have a scanner so it's digital documents only for the moment for me as well!
Paperless-ngx is so cool, I've been using it for a little while now myself but only for digital documents for now. But I recently migrated my setup to a local system and with that I explored the email processing feature a little more and I came up with a pretty nice little automated workflow.
Yeah to be fair I don't have a scanner so it's digital documents only for the moment for me as well!
This was neat, I hadn't heard of paperless-ngx before. Particularly the ability to ingest from an email seems interesting. I've had this idea in my head for a while to set up some automation to...
This was neat, I hadn't heard of paperless-ngx before. Particularly the ability to ingest from an email seems interesting. I've had this idea in my head for a while to set up some automation to read receipts from an email and create transactions in my finance tracking app, this might be a good starting point!
We’re maybe a couple years away from agentic workflows being given away for free like this, but keep an eye out. Plenty of companies are creating and selling their agentic tools.
We’re maybe a couple years away from agentic workflows being given away for free like this, but keep an eye out. Plenty of companies are creating and selling their agentic tools.
I've seen a prototype of an LLM setup that effectively does this. Send a receipt or invoice file to it via an API, and it can return a JSON object with all the line items and other info. This...
automation to read receipts from an email and create transactions in my finance tracking app
I've seen a prototype of an LLM setup that effectively does this. Send a receipt or invoice file to it via an API, and it can return a JSON object with all the line items and other info.
This would be a little difficult to setup purely from email, as more and more online stores only give you a link to the receipt in an email, and not a list of items purchased directly in the email body.
I recently set up paperless NGX in my house and I love it. It was really the software that tipped me over from overstressing my NAS running some services into just buying a dedicated machine to...
I recently set up paperless NGX in my house and I love it. It was really the software that tipped me over from overstressing my NAS running some services into just buying a dedicated machine to act as a server. The paperless NGX site itself is a really great resource with a lot of ideas for archiving and troubleshooting.
One of the issues I ran into at first was that I had my NAS mounted via the FSTAB over samba. Due to this paperless can't pick up file change events so it won't notice when you add things into the consume folder. There is something you can add to the compose or env that will instead periodically scan the consume folder on a timer. I've been using this with great success on a 10 min timer as I'm not really scanning too often and don't mind a delay.
Also the samba mount I think tends to throw off the export script. It wasn't working right out of the box for me, and one downside they mentioned was it needing to import to the same paperless version as well, so I ultimately just decided to run the backups late at night when I know I'm not scanning documents on the files. It has the actual files and the meta info, so I feel like it would still be pretty recoverable in the event of a loss.
Another factor I didn't think of during this whole setup was how important your choice of scanner is. The paperless site has a list of scanners that have some way to scan to the network and I ended up grabbing a brother ADS-1800W. I think it's a good scanner but did notice it has trouble separating the pages and sometimes pulls in 2 or 3 at once. I later saw this mentioned in reviews as and issue, and if I had thought about how much I interact with the scanner I'd have spent a little bit more or done more research on it.
But the point is right now the only interaction I have with paperless now that I'm set up is via the scanner. We have a wall mounted basket with PAY, SCAN, HOLD, and then below a little bit with SHRED next to the shredder. New mail that is a bill we put in pay, everything else to scan, a few we like to hold onto temporarily just for my wife's peace of mind and then either shred then or later in the shredder next to it. When the scanner pulls in two pages I have to scan multiple times and then it makes my sorting a bit more annoying later on down the line.
The sorting is honestly pleasant, it tags all new stuff with an inbox tag and I just work through it, as I'm adding in repeated bills and things it is doing well filling out most of the fields I want and I don't have to type too much. Honestly just takes me 10 min or so a week to keep it ordered.
So overall it's really great, I wasn't ready to start using it until I had all the pieces, the consume folder, a network scanner, a shredder, and tested backups. This thing is so much better than the mountains of paper we couldn't seem to manage before, I absolutely love it, and maybe it's a factor of my wife and I being unable to part with stuff, it's so much easier to digitally hoard it than physically.
I enjoyed this, so thanks for putting it together. I never thought about looking for someone like this, but I definitely want to organize my soft document pipeline, because right now it's ten...
I enjoyed this, so thanks for putting it together. I never thought about looking for someone like this, but I definitely want to organize my soft document pipeline, because right now it's ten different folders with ten different organization strategies, all containing at least some important stuff.
Some mild feedback on the article, if you're looking for it (if you're not QUICK PUSH BACK NOW):
I spent a lot of the article curious what portai er is but not enough to click out. You mention it a lot - a summary of what it is (or maybe the entire stack) would be a good opener.
Same thought about syncthing, but lesser since I can kind of assume what it is and what it does
I wanted you to elaborate on this: "Setting up the frontend for uploading via drag-and-drop is not trivial guys! This is speaking from experience."
Did you happen to look into options for ingesting straight from an app? That's where I always get caught. I want a way to send my files somewhere to be ingested, like you set up with syncthing > consumer folder > paperless, but using my phone. I've previously tried using Google drive as an intermediary which kinda works, but obviously has a lot of drawbacks.
Hey no worries. And I appreciate the warning before the feedback hah! Yeah. Sometimes I don't elaborate thinking no one's gonna read it. But you and the rest of the commenters have proven me...
Hey no worries. And I appreciate the warning before the feedback hah!
Yeah. Sometimes I don't elaborate thinking no one's gonna read it. But you and the rest of the commenters have proven me wrong.
I spent a lot of the article curious what portai er is but not enough to click out. You mention it a lot - a summary of what it is (or maybe the entire stack) would be a good opener.
Yeah, Portainer for me is a docker-compose wrapper. I'm sure it's more than that but that's my main use case. It has a web app so that means I don't have to login to my server and type up a bunch of magic spells in bash to do basic Docker container stuff.
Same thought about syncthing, but lesser since I can kind of assume what it is and what it does
Yeah. It's super handy. I recommend you to try it.
I wanted you to elaborate on this: "Setting up the frontend for uploading via drag-and-drop is not trivial guys! This is speaking from experience."
It was just that I encountered a discrepancy between browsers. We had a nasty bug that only happened in Chrome (at least from our perspective). It turns out that Chrome groups drag-and-dropped files into groups of 100 and we had to add code to go through the pages whereas Firefox and Safari just dumped them all into one big group (which we assumed was the case for all).
On top of that, there are the classic things that can go wrong when uploading big files. Network latency, failed uploads, the user leaving the page before everything is uploaded, "resuming" file uploads. The works.
Did you happen to look into options for ingesting straight from an app? That's where I always get caught.
Right off the top of my head you have at least two options:
Download the Paperless iOS app (I presume someone would have made one for Android) and "share" your PDFs to that app.
Setup an email to work with Paperless, like I did. If you have Gmail then you could do username+paperless@gmail.com and then use that to filter. From your phone, you send your PDFs and documents to that email and then you're golden.
Damn I guess we’re giving our RAG projects away for free now. Wild. I worked at a startup a while back whos entire product was a rag model just like this. They failed when the interest rates rose...
Damn I guess we’re giving our RAG projects away for free now. Wild. I worked at a startup a while back whos entire product was a rag model just like this. They failed when the interest rates rose and the vc funding dried up.
Could I set the consume folder inside my iCloud directory, and then just drop files to that directory (like from my phone) and have Paperless catch it on my Mac to process?
Could I set the consume folder inside my iCloud directory, and then just drop files to that directory (like from my phone) and have Paperless catch it on my Mac to process?
Hi, after a long hiatus (life got busy) I've finally had something new to share from tinkering with open-source projects.
Hope it could be useful for the community.
PS: I hope it's still okay that I'm posting my own blog here. I checked and I think I'm good. But if I'm mistaken, please do let me know if I should stop. :)
I was curious about your set up. I'm in the process of putting together a server for my home as well. Do you use a NAS and point your paperless instance towards that, or do you use storage on the server itself, or something else? I've been a bit paralyzed by choice regarding this, and think I'm leaning towards a "all in one" setup using one machine. Maybe building a NAS later would be worth it for me, maybe not!
For now I use a VPS service for both storage and hosting. I do want to get around to hosting things locally at some point although I am in the same boat as you with the analysis paralysis.
Maybe someone knowledgeable ends up finding this call for help ahah.
Thank you! I am still waffling between an all in one solution and a NAS + Server setup. I have the impression that separating out the functions is a good move but since my needs are so simple an all in one is sufficient
FWIW I'll put in a vote for Unraid. I set it up a couple years ago just on an old desktop and it's slowly been growing ever since. The big selling point to me was it doesn't need all same size hard drives. I've even moved it into a dedicated server case and upgraded the hardware and all moving the 'server' took was plugging in the same thumb drive. It's been great having the option to just throw up a docker and try something out. (I have paperless-ngx setup too and it's very nice, though I'm pretty lightweight with it.)
TrueNAS and alternatives sound neat but they're a bigger initial lift.
Not sure what you're looking to use it for, but unless you're expecting a several people to use it simultaneously it's not necessary to have dedicated machines for every piece. Useful for fun and redundancy though of course. I have a Pi5 as a second basket for things that are annoying to go down, pihole/syncthing/home assistant basically.
This sounds great. I already have an RPi4 running Adguard and synching. Never got into HA, but have dreamt about it. I think setting up a dedicated unRAID machine would be very nice. Just making sure I understand: do you have the unRAID machine working as just a NAS, or also hosting your services? I basically just want a few simple things running. It would only be me, and maybe 1 other person ever using it, so multiple machines is almost certainly unnecessary.
Edit: Services I would want are paperless, immich, and jellyfin, with other minor things
It's operating as both a NAS and general host. I could split the two up but it's really nice to be able to have effectively unlimited storage for whatever little dockers I want to play with since I have TB to spare.
I have a long list at this point but it's grown as I try something out for a bit and get it stable and then hear about something and add more.
I've got Plex as my main media sharing platform because it's easier to setup for family but I've spun up Jellyfin at the same time to test it out. Immich has been working great for the last few months, though I haven't pulled the trigger on deactivating Google Photos, mostly out of laziness. Also have calibre-web-automated and audiobookshelf for their respective formats. Nextcloud is nice for a self-hosted cloud replacement. Traccar is one I've been messing with because I really like using Google Maps to see where I've been. So far it's fine but I need to tweak it more and see about importing historical data. FreshRSS is nice to keep up with webcomics in a more organized manner. SearXNG is the engine the Kagi search engine uses, but for free. Heimdall is a nice homepage where I can put 'bookmarks' so I don't have to remember port numbers for all my services. There are a few game servers idling in the background, corekeeper/minecraft/etc. I also have ArchiveTeam-Warrior and Boinc dockers going to give me the warm and fuzzies.
Are you averse to using cloud storage? That's my plan, but wondering if I'm not considering all the things. I figure with encryption using my own key, I don't really care if AWS has my data, and would avoid the hassle of having to manage my own hardware and backups.
I think I am averse to cloud data, certainly more than most at baseline. I only want to upload my photos and documents to a service after they have been encrypted with something like cryptomator, Borg, etc. it is a more extreme position, so I understand it's not for everyone. I think if it's not on my own hardware, I don't own it!
I went with cloud storage just to get started. It is a pain to encrypt everything so I one day (hopefully within the year) wish to move to my own on-prem server for my apps.
Maybe the what option is on premise live storage with encrypted backups sent to cloud for redundamcy. Cold storage is pretty cheap and should your house ever burn down...
FWIW, I have a Synology NAS and I love it. It mostly self-manages, it runs Docker containers, and you can set them to be accessible over the web. I bet it would be great for something like this.
I've thought about going with an off-the-shelf solution but am actually excited to build something myself. It's been a while since I've built a computer and I used to enjoy it a good bit. I have a few drives hanging around that I could use in a selfbuild that might not be Synology compatible. Thank you for the recommendation though, these solutions have gotten very good!
Great article!
Paperless-ngx is so cool, I've been using it for a little while now myself but only for digital documents for now. But I recently migrated my setup to a local system and with that I explored the email processing feature a little more and I came up with a pretty nice little automated workflow.
The first thing I did was I created a dedicated email address for Paperless, just so I don't have to worry about anything getting imported that shouldn't. And then I made the Mail Rule in Paperless check for only emails coming from my personal email address with PDF attachments. This is pretty standard I think for automated mail importing.
But then in my mail client of choice, Thunderbird, I added this message filter. So when I archive an email and it has an attachment, and the sender is in my "Paperless Correspondents" address book in Thunderbird, it'll forward that email to the Paperless address and get automatically imported. Isn't that so cool!! I found it so neat when I figured out Thunderbird can do all this in an easy way.
I also run my setup with the consume directory available via Samba so should I want to bulk import something it's easy enough to do it there. Though the web UI works too in limited numbers.
Yeah to be fair I don't have a scanner so it's digital documents only for the moment for me as well!
This was neat, I hadn't heard of paperless-ngx before. Particularly the ability to ingest from an email seems interesting. I've had this idea in my head for a while to set up some automation to read receipts from an email and create transactions in my finance tracking app, this might be a good starting point!
We’re maybe a couple years away from agentic workflows being given away for free like this, but keep an eye out. Plenty of companies are creating and selling their agentic tools.
I've seen a prototype of an LLM setup that effectively does this. Send a receipt or invoice file to it via an API, and it can return a JSON object with all the line items and other info.
This would be a little difficult to setup purely from email, as more and more online stores only give you a link to the receipt in an email, and not a list of items purchased directly in the email body.
I'm sure it's just around the corner. Reading receipts automatically has been possible with my previous company's portal for reimbursements.
I recently set up paperless NGX in my house and I love it. It was really the software that tipped me over from overstressing my NAS running some services into just buying a dedicated machine to act as a server. The paperless NGX site itself is a really great resource with a lot of ideas for archiving and troubleshooting.
One of the issues I ran into at first was that I had my NAS mounted via the FSTAB over samba. Due to this paperless can't pick up file change events so it won't notice when you add things into the consume folder. There is something you can add to the compose or env that will instead periodically scan the consume folder on a timer. I've been using this with great success on a 10 min timer as I'm not really scanning too often and don't mind a delay.
Also the samba mount I think tends to throw off the export script. It wasn't working right out of the box for me, and one downside they mentioned was it needing to import to the same paperless version as well, so I ultimately just decided to run the backups late at night when I know I'm not scanning documents on the files. It has the actual files and the meta info, so I feel like it would still be pretty recoverable in the event of a loss.
Another factor I didn't think of during this whole setup was how important your choice of scanner is. The paperless site has a list of scanners that have some way to scan to the network and I ended up grabbing a brother ADS-1800W. I think it's a good scanner but did notice it has trouble separating the pages and sometimes pulls in 2 or 3 at once. I later saw this mentioned in reviews as and issue, and if I had thought about how much I interact with the scanner I'd have spent a little bit more or done more research on it.
But the point is right now the only interaction I have with paperless now that I'm set up is via the scanner. We have a wall mounted basket with PAY, SCAN, HOLD, and then below a little bit with SHRED next to the shredder. New mail that is a bill we put in pay, everything else to scan, a few we like to hold onto temporarily just for my wife's peace of mind and then either shred then or later in the shredder next to it. When the scanner pulls in two pages I have to scan multiple times and then it makes my sorting a bit more annoying later on down the line.
The sorting is honestly pleasant, it tags all new stuff with an inbox tag and I just work through it, as I'm adding in repeated bills and things it is doing well filling out most of the fields I want and I don't have to type too much. Honestly just takes me 10 min or so a week to keep it ordered.
So overall it's really great, I wasn't ready to start using it until I had all the pieces, the consume folder, a network scanner, a shredder, and tested backups. This thing is so much better than the mountains of paper we couldn't seem to manage before, I absolutely love it, and maybe it's a factor of my wife and I being unable to part with stuff, it's so much easier to digitally hoard it than physically.
Thank you for the comment. Happy to hear that this app has motivated you to take the plunge and get a NAS. Hope I'll get to that too at some point!
I enjoyed this, so thanks for putting it together. I never thought about looking for someone like this, but I definitely want to organize my soft document pipeline, because right now it's ten different folders with ten different organization strategies, all containing at least some important stuff.
Some mild feedback on the article, if you're looking for it (if you're not QUICK PUSH BACK NOW):
Did you happen to look into options for ingesting straight from an app? That's where I always get caught. I want a way to send my files somewhere to be ingested, like you set up with syncthing > consumer folder > paperless, but using my phone. I've previously tried using Google drive as an intermediary which kinda works, but obviously has a lot of drawbacks.
Hey no worries. And I appreciate the warning before the feedback hah!
Yeah. Sometimes I don't elaborate thinking no one's gonna read it. But you and the rest of the commenters have proven me wrong.
Yeah, Portainer for me is a docker-compose wrapper. I'm sure it's more than that but that's my main use case. It has a web app so that means I don't have to login to my server and type up a bunch of magic spells in bash to do basic Docker container stuff.
Yeah. It's super handy. I recommend you to try it.
It was just that I encountered a discrepancy between browsers. We had a nasty bug that only happened in Chrome (at least from our perspective). It turns out that Chrome groups drag-and-dropped files into groups of 100 and we had to add code to go through the pages whereas Firefox and Safari just dumped them all into one big group (which we assumed was the case for all).
On top of that, there are the classic things that can go wrong when uploading big files. Network latency, failed uploads, the user leaving the page before everything is uploaded, "resuming" file uploads. The works.
Right off the top of my head you have at least two options:
Download the Paperless iOS app (I presume someone would have made one for Android) and "share" your PDFs to that app.
Setup an email to work with Paperless, like I did. If you have Gmail then you could do username+paperless@gmail.com and then use that to filter. From your phone, you send your PDFs and documents to that email and then you're golden.
Thanks so much for responding and the extra info. I'm definitely inspired and having your recipe as guidance will get me started more quickly.
Damn I guess we’re giving our RAG projects away for free now. Wild. I worked at a startup a while back whos entire product was a rag model just like this. They failed when the interest rates rose and the vc funding dried up.
Could I set the
consumefolder inside my iCloud directory, and then just drop files to that directory (like from my phone) and have Paperless catch it on my Mac to process?