jensabacik

Thanks! The magic is in the implementation. You could replicate many of the features and sometimes get similar result, but in practice it would get too rigid, very expensive and often the result would not be nearly as good.

Memory couldn't be replicated. ChatGPT did a good improvement to long term memory recently, but it's not getting close to Good Assistant. And memory affects every other feature.

Other features like notes & tasks integration, reminders etc. could be replicated with custom setup, the limitation there would be what I talked about in the post - the generic agent does what you tell it to do, nothing more. This means that it can't be "aware" of your context, notes, tasks, goals etc, unless you first ask it to read them, in the beginning of each conversation or in each automation.

Does this make more sense? :-)

This app is my passion project. For a while I had an idea of human-like assistant I would use to bring structure, necessary planning and motivation to me pursuing the goals I had. Things like learning a language and taking my startup to a new phase. It felt like AI is there in terms of capability, but product like ChatGPT went in different direction. Which is good for me because I got excited about building it myself :-)

I started working on Good Assistant in the beginning of 2025, released first version in the beginning of 2026, and working until now to bring it to the level of usefulness I was imagining from the start.

What does it do?

An app with tasks, notes, long term goals, and chat. Your assistant helps you establish concrete plan for pursuing your few selected goals, helps with anything AI can help with, and it's proactive. Every morning it sends you a message with suggested plan for the day which you can adjust and commit to. It's aware of your schedule and work - one of the aims is to help you plan in the way that's not overwhelming, even when tasks piled up. The assistant also sends you a message during the day, from time to time, whenever it's helpful based on the context.

I identified a few areas where current AI apps are lacking, where I thought a much better experience is possible. And developed in that direction. The result is that the assistant is unique in a few ways, better than any AI product I tried.

Interesting bits:

Memory. Chat apps like Claude and ChatGPT are improving, but Good Assistant has the best memory system. There are many constantly maintained layers that make it possible, static and dynamic, and the result is that the interaction feels often closer to talking with human than with an AI. This is also what makes the assistant expensive to run (and costs $29 per month). The rich memory accounts for majority of tokens used and needs background tasks that build up the "static" parts daily. But the result is an assistant that remembers what you mentioned once 1 year ago, naturally, without asking about it.

Integration of tasks, notes, goals, reminders, calendar. These features are already possible to use in your favorite AI, via different MCP connections. The problem is that in practice, you need to instruct the AI to use it. It doesn't naturally read your tasks and notes and uses that knowledge to be more helpful. But Good Assistant does because the integration of these features is woven into the app on all levels.

I welcome you to try your own assistant, there is a 1 week free trial, and tell me what you think about the idea. I am well aware of an AI assistant being already overcrowded space. This one is built to solve one particular problem (making real sustained progress on goals) better than any other tool.

Nice idea! And I see that there's also automated transcript for SEO & accessibility. Is there some standardised way to upload the content, or any image goes?

Hi Saai, where can I find any info about the company & people behind the project? Terms and other docs on the page don't say much, About page is not about the company etc.

The project looks very impressive!

Creative take, I like the visual representation on infinite canvas a lot. Feels like freedom. What doesn't feel like freedom nor future is the manual selection of tools. AI should naturally pick the right tools out of all possible tools, choosing what abilities or data it can have each time is hassle. What is the reasoning for manual and explicit tools? Still, very cool UI :-)

Nice implementation. How do operating systems handle this by default? Do they stick to the last selected WiFi no matter the speed and signal strength, even if there are saved networks available?

Also, are you thinking of making this into an app that would selected and switch networks automatically? Like a Mac menubar app.

It looks very impressive, I might try it at work. We have a challenge of making our own AI chat "integrated" as an MCP in any compatible tool like Claude.

Very cool idea. You don't plan to add data for Europe any time soon?

What was your initial problem? I usually work with contexts where they were created, like code session in Cursor / chat in Claude app. When do you have the need to often transfer the history?