I Built a Python script that uses a local Ollama LLM to automatically find and add movies to Radarr.
It picks random films from your library, asks Ollama for similar suggestions based on theme and atmosphere, validates against OMDb, scores with plot embeddings, then adds the top results to Radarr automatically.
Examples:
- Whiplash → La La Land, Birdman, All That Jazz
- The Thing → In the Mouth of Madness, It Follows, The Descent
- In Bruges → Seven Psychopaths, Dead Man’s Shoes
Features:
- 100% local, no external AI API
- –auto mode for daily cron/Task Scheduler
- –genre “Horror” for themed movie nights
- Persistent blacklist, configurable quality profile
- Works on Windows, Linux, Mac
GitHub: https://github.com/nikodindon/radarr-movie-recommender
Did you build it, though, or did Claude code it?
Built with Claude by the looks of things. Not sure if Claude was used to generate the boilerplate and whether the dev reviewed it after or whether Claude did all of it, but definitely Claude was used for some of it. I recognise the coding style that Claude outputs and the bugs that it implements that will cause TypeErrors if not handled.
FWIW, I’m not against using AI as an assistant for coding (I do it too, using Claude and Vercel as assistants) just as long as the code is reviewed and understood in full by the dev before publishing.
FWIW, I’m not against using AI as an assistant for coding (I do it too, using Claude and Vercel as assistants) just as long as the code is reviewed and understood in full* by the dev before publishing. *my emphasis
A very sane take. I do wish devs would fully disclose this on their github or other. That way, if the project is seasoned, well starred, et al, and the dev used AI as an assistant, then the user gets to decide. Given all the criteria are met, I would deploy it.
I will say that I have observed what seems like a pretty decent up tick in selfhosted apps, and I would be willing to bet a goodly amount of them have at the very least, used AI in some capacity, if not most/all code. I don’t have any solid evidence to back that up but it just seems that way to me.
Since no one is leaving critical comments that might explain all the downvotes, I’m going to assume they’re reflexively anti-AI, which frankly, is a position that I’m sympathetic to.
But one of the benign useful things I already use AI for, is giving it criterias for shows and asking it to generate lists.
So I think your project is pretty neat and well within the scope of actually useful things that AI models, especially local ones, can provide the users.
LLMs are not the tool for a recommender job
No LLM use is benign. The effects on the environment, the internet, and society are real, and that cannot be ignored.
You can make the argument that in some cases it is justified, e.g.: for scientific research.
Do You drive a car? Eat meat? Fly for holidays?
Nothing is black and white.
Saw it was already commented about CO2, so I thought I’d counter-point your environment claim regarding water usage (since that is something I’ve seen a lot of too).
The ISSA had a call to action due to the AI water use “crisis”: https://www.issa.com/industry-news/ai-data-center-water-consumption-is-creating-an-unprecedented-crisis-in-the-united-states/
68 billion gallons of water by 2028! That’s a lot…right? Well, what I found is that this is somewhat of a bad faith argument. 68 billion gallons annually is a lot for one town, but those are numbers from a national level and it isn’t compared to usage from anything else. So, lets look at US agriculture (that’s something that’s tracked very well by the USDA): https://www.nass.usda.gov/Publications/Highlights/2024/Census22_HL_Irrigation_4.pdf
That’s 26.4 trillion gallons of water annually. So, AI datacenter represents 0.26% of agriculture consumption. If AI datacenter consumption is a crisis, why is agriculture consumption not a crisis? You could argue that agriculture produces “something useful”, but usefulness doesn’t factor into the scarcity of a resource. So, either its not a crisis, or you are cherry picking something that has no meaningful outcome to solving the problem.
yeah, I think the whole “water” argument really dilutes the case against data centers.
On a serious note, the argument works for areas that already struggle to supply enough water for consumers. Otherwise, we should be focusing more on the power stress to the grid, and the domino effect on supply chain of hardware cost increases that it’s happening across many industries. It started with GPUs, now it’s CPU, storage, networking equipment, and other components.
If these prices are too high for a couple of years, we’ll start seeing generalized price increases as companies need to pass along the costs to consumers.
I think the supply chain issue is probably the most pressing out of all of them. The other points people have are either non-issues or a result of dropping usage hogs into existing electrical infrastructure. Infrastructure can be updated, though.
Supply chain is different. There isn’t a supply shortage of chips, its that profitability dictates you should sell them to datacenters or adjacent industry. Unlike infrastructure where you can just build out more, adding more supply for chips just means you have more to sell to datacenters. Since the demand is there, end of day profits will always win.
The effects on the environment
Didn’t down vote you. I hear this line of complaint in conjunction with AI, especially if the person saying it is anti-AI. Without even calculating in AI, some 25 million metric tons of CO2 emissions annually from streaming and content consumption. Computers, smartphones, and tablets can emit around 200 million metric tons CO2 per year in electrical consumption. Take data centers for instance. If they are powered by fossil fuels, this can add about 100 million metric tons of CO2 emissions. Infrastructure contributes around 50 million metric tons of CO2 per year.
Now…who wants to turn off their servers and computers? Volunteers? While it is true that AI does contribute, we’re already pumping out some significant CO2 without it. Until we start switching to renewable energy globally, this will continue to climb with or without AI. It seems tho, that we will have to deplete the fossil fuel supply globally before renewables become the de facto standard.
chill, this is extracting text embeddings from a local model, not generating feature-length films
that’s like saying “no jet use is benign” meant for comparing a private jet to a jet-ski
the generative aspect is not even used here
So running a local model is unforgivable, but “scientific research” running on hyperscalers, can be justified?
Huh? There are other ways to link similarities of movies without the use of a llm. You may use ai to find similar movies but it’s nonsense that everyone has to ask a llm to link movies.
no one is saying everyone has to ask an LLM for movie recommendations
OP wrote a python script that call a llm to ask for a recommendation.
But you are right, op doesn’t say that everyone shall do it
No, it also doesn’t do that. It gets embeddings from an LLM and uses that to rank candidates.
Are you a trollm?
If not, I’m just too stupid to understand op.
I Built a Python script that uses a local Ollama LLM to automatically find and add movies to Radarr.
OP wrote a python script that call a llm to ask for a recommendation.
If that’s not the same, I don’t know what is. Gotta go back to school, I guess.
It’s not, I read the code. It’s not merely asking the LLM for recommendations, it’s using embeddings to compute scores based on similarities.
It’s a lot closer to a more traditional natural language processing than to how my dad would use GPT to discuss philosophy.
I love how you actually did something instead of jumping on the generalization train. I hate that about Lemmy right now.
that’s pretty cool, this is just the wrong crowd, don’t worry about the downvotes
I remember building something vaguely related in a university course on AI before ChatGPT was released and the whole LLM thing hadn’t taken off.
The user had the option to enter a couple movies (so long as they were present in the weird semantic database thing our professor told us to use) and we calculated a similarity matrix between them and all other movies in the database based on their tags and by putting the description through a natural language processing pipeline.
The result was the user getting a couple surprisingly accurate recommendations.
Considering we had to calculate this similarity score for every movie in the database it was obviously not very efficient but I wonder how it would scale up against current LLM models, both in terms of accuracy and energy efficiency.
One issue, if you want to call it that, is that our approach was deterministic. Enter the same movies, get the same results. I don’t think an LLM is as predictable for that
I’m not an expert, but LLMs should still be deterministic. If you run the model with 0 creativity (or whatever the randomness setting is called) and provide exactly the same input, it should provide the same output. That’s not how it’s usually configured, but it should be possible. Now, if you change the input at all (change order of movies, misspell a title, etc) then the output can change in an unpredictable way
Yes. I think determinism a misunderstood concept. In computing, it means exact same input leads to always the same output. Could be entirely wrong, though. As long as it stays the same. There’s some benefit in introducing randomness to AI. But it can be run in an entirely deterministic way as well. Just depends on the settings. (It’s called “temperature”.)
One issue, if you want to call it that, is that our approach was deterministic. Enter the same movies, get the same results. I don’t think an LLM is as predictable for that
Maybe lowering the temperature will help with this?
Besides, a tinge of randomness could even be considered a fun feature.
Y’all need LLMs to tell you what to watch, now?
Screw AI.
A recommendation for Moonrise Kingdom based on Mickey 17? The genres might match, but those are totally different movies.
Also, A Bug’s Life from Mickey 17!?
How does this compare to an ML approach?
are you training or just using an LLM for this?
There’s no training, the LLM embeddings are used to compare the plots via a cosine similarity, then a simple weighted score with other data sources is used to rank the candidates. There’s no training, evaluation, or ground-truth, it’s just a simple tool to start using.
A bugs life from mickey 17?
Explain OP
Bugs, obviously.
