@swiftie2996 Instead of Twilo you can leverage telegram.. if that is for a personal project...
Twilo cost you money even for the mvp!
You can integrate twilo later..
And i would suggest.. leverage the frameworks and try to invest your time in the core harness engineering..
@swiftie2996 Hmmm, have a question
Why have you chosen the base llama.cpp raw, instead of the existing frameworks like langchain or langflow?
If that is an agent it has to use some tools, right!
As you said no mcp, do you code those tools by yourself?
@swiftie2996@GithubProjects I would like to see that!
Does it have an interface like a desktop app or just the harness?
Is it your product or open-source project?
Good luck buddy!
I think both needed...
cuz animations leads to better experience and information must be planned.. have to track each phrase...
and after designing, the next hardest part is what information should we put on the website, also where & how much!
As a end user what do you wants to see in a website in 2026!
- Animations
- Information
- Both but limited (more content focused)
** in this era animations matters too along with the content ig...
As a end user what do you wants to see in a website in 2026!
- Animations
- Information
- Both but limited (more content focused)
** in this era animations matters too along with the content ig...
Deep Researcher v2 is hitting the brutal local-model wall 😵💫!
8k context on Ollama is destroying reasoning quality. One query can touch 40+ pages, while full research needs 300-400 pages processed.
Need better memory architecture + stronger hardware... Ig..
Working on it!!
@LearnWhyWeTrust Hmm... it's quite deterministic....
If you have a name then you get popular quick...
If you don't have some good dp "may" helps..
I never got this lucky tho tried lots of things...
pixelImage works so beautifully... I am realizing now that this is the best software that I've built so far.. It's a simple app but useful for me as a developer...
✌️🏻...
The visualization in chatGPT is pretty good, actually!... I am thinking of adding something like this in "Deep Researcher v3"..
It just simplifies things more than text.