100% Local Super Easy Private Email RAG Setup | Ollama - Gmail ++

100% Local Super Easy Email RAG Setup | Ollama - Gmail ++

🔥 GitHub Repo:

👊 Become a member and get access to GitHub and Code:

🤖 Great AI Engineer Course:

📧 Join the newsletter:

🌐 My website:

Members GH Repo:

In this video I create a 100% local Email RAG with gmail and outlook support, search and talk your private email 100% locally using Ollama

00:00 Local Email RAG Intro
03:04 Local Email RAG Code 1
04:38 Gmail / Outlook Setup
06:10 Local Email RAG Code 2
08:35 Testing the Email Rag Syste
10:42 All About AI Community

[BThunder]

Sweet! I want my local AI to look at all my emails all the time, in real time. Key words to pull in summary as you show here but with seamless simultaneous option for questions like you show here. I would use it all day long at work.

[jsgordon]

Thanks for all your hard work!

It would be super helpful to flash the end result at the beginning (or fly up window) . While this video led to what I understood to be the intended goal, sometimes I'm wrong or don't know exactly going in. I try to imagine the end sometimes, but I'm as likely to be inaccurate the earlier I try guessing or making up the end presentation layer (far less important than everything beneath -- like a cars paint, sure we see it first, but it's contribution to [function car] is very low percent.)

Keep up great work. I enjoy your knowledge and curiosity and humility. =={Worthwhile, joyous, relatable}

[thecoolbeanz]

Kris, man this is excellent work! Thank you for all your efforts put into these repos and videos. I've been watching pretty much since ChatGPT was first released and you were just starting out (or just getting lots of subscribers).

I notice in your PowerShell (PS) that when you type your Python file commands, for example 'python collect_emails.py', your PS then provides an autocomplete suggestion for the Python file's internal arguments "--startdate --enddate".

I was wondering how you set it up to do this? I've tried reading some online docs but no luck so far. If you could point me in the right direction it would be appreciated.

I'm working on a local LLM to be my business admin assistant currently. Your videos are helpful.

[settlece]

how handy is this it must be everybody's problem with the reams and reams of emails we've got
thanks for all your work awesome





i'm not gonna talk about taylor swift 8)

[Enkumnu]

Excellent. Very good job. What do you think to integrate a Vector DB ?

[maksimandronenkov]

Thanks for the content. Looks cool.
Any ideas how to use email script only with one mail system, for example Gmail only?
If I specify only Gmail account, script also wants to have also Outlook credentials.

[FrederickMbuya]

Totally new so excuse if nonsensical question, but would it be possible to pull in several years worth of email and store for interrogation

[FilmFactry]

I tried doing this awhile ago with Bard no luck. This is pulling from the cloud. I've downloaded a selection of emails using Google Takeout and they are downloaded into a MBOX file locally. Would that not be a good way to work with the data? Also. How large of an email dataset can you work with? Thanks!

[yagoa]

is there any easy way of creating extra context for rag embeddings using llms?

[yagoa]

fyi you did not blur enough to hide the text

[supra]

I'm having an issue connecting to outlook. Using outlook 365. I get an error in line 141 and 153 in collect emails. And line 202, 312, 1306, 1310, 1319 in imaplib. Does anyone have a solution? I'm new to running programs so don't have too much to try