PRIVATE alternative to ChatGPT (free). Work with PDF, CSV, DOCX and code

[vlad_x]

Just recommended, downloading for testing...
A completely PRIVATE alternative to ChatGPT has been released — AnythingLLM keeps your requests and data secret from everyone.
The service can work with PDF, CSV, DOCX and code. On its basis, you can build AI agents for automation, analysis and collection of information.
The best part is that the neural network is not inferior in power to ChatGPT.


Free for local usage

 

[Correctively Objective]

Also take a look at TypingMind. I'm not affiliated in any way to them. The product is just awesome.

 

[sudhirnyc]

Just recommended, downloading for testing...
A completely PRIVATE alternative to ChatGPT has been released — AnythingLLM keeps your requests and data secret from everyone.
The service can work with PDF, CSV, DOCX and code. On its basis, you can build AI agents for automation, analysis and collection of information.
The best part is that the neural network is not inferior in power to ChatGPT.


Free for local usage

Thanks for sharing, would be using this extensively.

 

[NoSoup4U]

Old information.
There is also ollama

 

[M4DM4X]

Just recommended, downloading for testing...
A completely PRIVATE alternative to ChatGPT has been released — AnythingLLM keeps your requests and data secret from everyone.
The service can work with PDF, CSV, DOCX and code. On its basis, you can build AI agents for automation, analysis and collection of information.
The best part is that the neural network is not inferior in power to ChatGPT.


Free for local usage

I knew this existed but I thought it was super complex to set up, thanks for mentioning this tool it seems very easy to use.
Do you know if content generation can be automated with it, like with openai api? And since it's local is there a risk of always getting the same kind of content?
Thanks mate

Old information.
There is also ollama

Thanks for sharing this info as well, is ollama capable of generating content locally to automate website creation in python?

 

[vlad_x]

I knew this existed but I thought it was super complex to set up, thanks for mentioning this tool it seems very easy to use.
Do you know if content generation can be automated with it, like with openai api? And since it's local is there a risk of always getting the same kind of content?
Thanks mate

Thanks for sharing this info as well, is ollama capable of generating content locally to automate website creation in python?

I just started dev content rewriting for a WordPress site. I hope to achieve good results.
AnythingLLM support RAG this is what we need for a good generations.
My plans include creating a content factory using n8n.
In addition to text, it will automatically generate videos and images for products.
The server infrastructure is ready, all that remains is to train the models and connect automation with WordPress.

Yes, ollama can theoretically, need to train the model.
If you need a text generation model look at https://github.com/oobabooga/text-generation-webui?tab=readme-ov-file also supported AnythingLLM.
But ollama do not support a RAG, that's why I don't consider this option. ollama is not a the same AnythingLLM
Thanks for the feedback.

 

[vlad_x]

And also, it needs to be tested, this LLM can be connected to the Cursor as a model for coding.

 

[Killswitchs]

looks cool on paper but let’s be real, chatgpt got way better training, speed, and stability. those private setups sound nice till you hit bugs, slow response, or weak context handling. and I am talking as someone who tried over 8+ cheaper alternatives to chatgpt

 

[vlad_x]

looks cool on paper but let’s be real, chatgpt got way better training, speed, and stability. those private setups sound nice till you hit bugs, slow response, or weak context handling. and I am talking as someone who tried over 8+ cheaper alternatives to chatgpt

I'm sure you're right.
But how do you work with confidential data that can't be leaked? We can't give this personal information to ChatGPT
We need to use a local model, but which one?

 

[Killswitchs]

I'm sure you're right.
But how do you work with confidential data that can't be leaked? We can't give this personal information to ChatGPT
We need to use a local model, but which one?

fair point if it’s real sensitive data, but most don’t actually need local unless they handling hardcore private stuff. if you really must go local, try llama3 with ollama or nous hermes fine tuned. just know you’ll sacrifice quality for control. your call.

 

[Spicy Chromosomes]

Decent option for people looking for privacy, in the end Gemini, ChatGPT and other mainstream AI tools are another surveilance tools for big techs

 

[vlad_x]

fair point if it’s real sensitive data, but most don’t actually need local unless they handling hardcore private stuff. if you really must go local, try llama3 with ollama or nous hermes fine tuned. just know you’ll sacrifice quality for control. your call.

Yes, I need to look into this in more detail and test it. I'm still confused.
Thanks

 

[flackz]

For local use, I would recommend the Gemma 3 E4B model from Google. It's lightweight and fast, and it can easily run on older GTX 16xx GPUs. Its quality is similar to that of GPT-4o-mini.

Use it for free with LM Studio.

 

[NoSoup4U]

I knew this existed but I thought it was super complex to set up, thanks for mentioning this tool it seems very easy to use.
Do you know if content generation can be automated with it, like with openai api? And since it's local is there a risk of always getting the same kind of content?
Thanks mate

Thanks for sharing this info as well, is ollama capable of generating content locally to automate website creation in python?

I don’t know if I would use anythingLLM for that use case. n8n automation platform or similar would yield better results.

You’ll also need to do a series of micro prompts. You’re likely asking too much as once. You also need to prompt engineer for outcomes and data you can process to create the site. Advanced prompt engineering and model testing is required for this.

Yes I do all of this. Only work to build industry worthy solutions to get anyone’s attentions.

 

[vlad_x]

I rewrite product descriptions.
Now there is an error when sending results to WooCommerce.
There's only a little time left until the result...

I don’t know if I would use anythingLLM for that use case. n8n automation platform or similar would yield better results.

You’ll also need to do a series of micro prompts. You’re likely asking too much as once. You also need to prompt engineer for outcomes and data you can process to create the site. Advanced prompt engineering and model testing is required for this.

Yes I do all of this. Only work to build industry worthy solutions to get anyone’s attentions.

 

[NoSoup4U]

I rewrite product descriptions.
Now there is an error when sending results to WooCommerce.
There's only a little time left until the result...


View attachment 461713View attachment 461714

DM me if you want professional help. We do troubleshooting and on call for advanced stuff and basic setup.

 

[vlad_x]

DM me if you want professional help. We do troubleshooting and on call for advanced stuff and basic setup.

Okay. I'd be glad to collaborate. I'm looking for a team.

 

[code API]

I have been using AnythingLLM for over three months already, not looking back.

My coding set up:
AnythingLLM and Notepad ++

 

[vlad_x]

I have been using AnythingLLM for over three months already, not looking back.

My coding set up:
AnythingLLM and Notepad ++

Can you recommend a free model for using a hybrid work API method for agent chat: your documents + web search?

 

[M4DM4X]

I don’t know if I would use anythingLLM for that use case. n8n automation platform or similar would yield better results.

You’ll also need to do a series of micro prompts. You’re likely asking too much as once. You also need to prompt engineer for outcomes and data you can process to create the site. Advanced prompt engineering and model testing is required for this.

Yes I do all of this. Only work to build industry worthy solutions to get anyone’s attentions.

Thanks for your reply.
I read one needs a powerful laptop to run models locally, with powerful graphic cards, vram etc, Dont have a laptop like this so I'll continue using the online version of openAI for my projects