Ultimate plan features brainstorming

oxlvlnle

fxgn While I second this, Kagi team is quick to deal with whatever technical problems. Just message in Discord and usually you will get an answer in a short time, no matter if you are ultimate user or not.

Thibaultmol

oxlvlnle I Third it BUT keep in mind that not everyone uses Discord. A lot of privacy minded folks aren't on Discord for example. So they'd ideally need to have a thing where ultimate subs get their emails high prio

MartyM

Hi. I would like to suggest adding to the AI assistant the ability to send files to the user, for example, word, excel, pdf documents. This is something I really miss in my daily work. 🙂

marcel001122

Give ultimate subscribers a way to become anonymous in Kagi ecosystem. Vlad mentioned in another thread that it is possible to do Mullvad style logins but if that comes with a cost. Lock that behind the Ultimate plan and you got a winner! Those that want anonymity get anonymity.. if there are many of those amongst us, Kagi gets an uptick in Ultimate users.

lucascantor

I'd really love a native mobile app for the assistant.

Even using it as a PWA on my iPhone is very clunky, with some of the assistant's UI elements obscured behind built-in iOS UI elements:

...and being forced to approve microphone access literally EVERY time I want to use it:

Bonus points if it could integrate directly with Siri, but I know that would take partnership with Apple, who thus far have only offered ChatGPT, though they've said that more partnerships will come...

LinuxLefty

I just signed up for the Ultimate plan and it's been great, but I was disappointed by the lack of image generation, as that's included in many other service's offerings. It would be great to have that feature added 🙂

RoxyRoxyRoxy

Not sure if it's been mentioned yet but an increased/uncapped limit to custom CSS would be great (if feasible technically ofc)

dps

Memory for pinned threads could be an ultimate only feature https://kagifeedback.org/d/5945-memory-for-assistant-like-chatgpt

goulot_situez

dps I'd like more than span options too: 24h (default), 7 days, 30 days, 1 year, Infinite.

LXIN

Below is not request as much as a concept that will become relevant once everyone starts to use automated web browsing.

0: Kagi has inexpensive hardware attestable minimum spec devices somewhere.
1: Ultimate enables you to gain access to a Linux privileged user in a device.
2: This happens such that first the device is attested against a Kagi talisman or DIY device, with attestation keypair & storage.
3: Attestation successful => Locally decrypt storage & tell you did to Kagi => Kagi elevates you to virtual non-attestable more powerful runtime. This still cannot run hardware wise LLM inference or do heavy computing. LLM runtime comes from providers.
4: Your local store information populates now Kagi's virtual runtime, that is for now online & SSH-able if you want, such that there is Kagi plain assistant discussions as one other account.
5: Each agentic custom assistant, e.g. normal discussion can be you spawned its own user, turning it agentic.
6: The Lens of that assistant determines its users permissions & Terminal Emulator tools available.
7: The assistant may now interact with the world by using the terminal.
8: An assistant may if given permission to message other agents of alter their permissions or create new asisstants.
9: Underlying all of this is a Kagi system-wids package.
10: For example, when creating a video game, a primary assistant could coordinate with specialized designer, coder, and playtester assistants. The workflow proceeds like this: designer creates concepts, passes them to main asistsant whhok passes them to coder, the coder for implementation, and the main assistant then forwards the result to the playtester for evaluation.

(speculative concept)

RoxyRoxyRoxy

LXIN Sooo.... cloud VM hosting with AI agents? Seems expensive and unsure what the usecase would be that couldn't be done (for example) with a desktop app capable of that same agenting running on one's own PC. Plus your tenth bullet point seems directly at odds with Kagi's stance on AI. Personally while I love the Assistant, I don't like the idea of some autonomous AI-powered Web, and don't want to see all human interaction and creativity replaced by LLMs and generative AI, even one powered by Kagi. You've got some interesting ideas here I do support (further agenting capabilities in the Assistant could be useful, not sure if that's coming in Ki) but idk about the rest.

LXIN

RoxyRoxyRoxy

Quick Answer

Brainstorming is a creative thinking technique used

to generate a large number of ideas quickly and collaboratively. It typically involves a group of people who interact to suggest ideas spontaneously in response to a specific prompt or problem. The primary goals of brainstorming are to encourage free thinking, foster creativity, and explore various solutions to a defined issue.

Key characteristics of brainstorming include:

Group Interaction: Participants share their thoughts and ideas without criticism, promoting an open environment.
Idea Generation: The focus is on quantity over quality, allowing for a wide range of suggestions.
Spontaneity: Ideas are generated in a free-flowing manner, often leading to unexpected insights.
Consensus Building: After generating ideas, the group can refine and merge them to find the best solutions12.
Overall, brainstorming is an effective method for teams to collaboratively tackle challenges and innovate34.

...

... speculative.

The comment was a shit also labelled as a one

I did not expect feedback on its quality, appreciate your comment, and don't intend to refine my shit so I might not reply to it more than to clarify that in the suggestion, LLMs would be provided by external providers, the storage controlled controlled physically by the user, and software itself run by Kagi, which somewhat albeit unintentionally and distantly resembles one vaporware / concept-stage design that spun off from QubesOS project; and that eventually, agency will become "a feature", terminal emulators being a natural & powerful extension of LLMs. However that's just to clarify the original meaning of it.

I think the issue here would however be more of a procedual or communication flaw in the forum than anything of yours or mine.

The feature request system has good design. Alas it seems to me that the brainstorming, consolidation & reviewing phases, going from really hastily sketched down ideas to "concepts that pass basic critical examination & Kagi says are plausible in principle", generating refined work, could use, forum cultural, usability & epistemological work.

What you want is something that doesn't let shit pass through but let's it be thrown around & systematically compared. Currently in kagifeedback.org shit is only allowed to be thrown which is a good thing as long as Kagi doesn't take "the feedback" too seriously (see why Asahi head developer resigned in 2025-02).

RoxyRoxyRoxy

LXIN Sorry I’m not sure why the hostility? Like I said you’ve got a few good ideas in there but the rest seems wildly out of scope and expensive for Kagi.
I’d also argue in your second point about the feedback process that while rough idea sharing is good, typically things should indeed pass “basic critical examination” before being posted; lest we be flooded with impossible posts such as “Kagi should make their own LLM and have it be better and faster than anyone else’s”.
Finally as a purveyor of internet drama I’d like to say that’s far from the reason the Asahi dev quit; he was told off by Linus Torvalds himself to stop attacking other maintainers on social media and bringing drama into the mailing lists.
Apologies if I’ve been too rude in my previous reply or this one; simply trying to figure out what you mean, and helping to make it something more in scope and feasible for Kagi.

« Previous Page