Marklet is an open source alt-AI client primarily intended for use with local models (i.e. in conjunction with Ollama or a similar runner). Its primary purpose is to do this simply and to do it well.
Marklet is a native desktop application which provides a near out-of-box experience, comprising everything except the models and runner. It is clean, principled and designed for those who do not want bloat or web-UIs.
It prioritises the individual, and is personal in the way a PC was once personal.
Above, early version 0.4 in dark mode.
UNDER ACTIVE DEVELOPMENT 2026
This is an preview release encompassing the primary aspects of the user interface and database. It does not currently contain the code needed to connect it to a model. There is no actual "AI" in current releases. It is coming. Click Watch for updates.
That being said, you should find the user interface very polished. You can download and run it, and it will generate stub response messages intended to exercise the user interface.
It initially supports:
- Linux
- Windows
Runable Linux AppImage and simple Windows zip are provided only for this release. It should build and run on macOS, but the means to test this is currently unavailable to the authors (feedback on this appreciated).
Marklet is a Sovereign AI initiative predicated on the idea that, while large AI services bring a kind of creative destruction, local AI will persist into the future. Moreover, local AI will grant agency to those who can maintain their own autonomy in the face of convenience and automation that AI appears to offer. As such, Marklet is not itself vibe-coded, but created with native thought. The intention is for it to be something a little different and go against trends as needed.
An approximate roadmap is as follows:
- Continued development of the user interface and data backend
- Completion of "note taking" interface
- Completion of "project" interface
- Implement connection to model runners and context generation
- Syntax high-lighting for fenced code
- LaTeX math symbols in markdown
- Support for attachments, images
- Text to speech, yet undecided features
- A plugin architecture
Marklet is not intended to support image or video generation capabilities. Any text-to-speech capability will be for accessibility only and not to mimic real human voices.
There are likely to be several interim releases before an initial version 1.0. Work continues.
The following principles, grounded in an understanding of entropy, guide the development of this software:
-
Machines Should Not Deceive
Chat and AI systems should not masquerade as human, nor use manipulative psychology such as responses designed to appeal to ego. Modern speech synthesis should contain some audible trait making it discernible from that of human. The same applies to simulated visual appearances and all modes of perception.
-
Automation of the Law of Unintended Consequences is a Bad Idea
Despite the allure, agentic automation and machine sub-goals are to be eschewed. Information is dissipated as deterministic processes interact with the environment. Without consciousness to intervene, consequences move away from desirable outcomes. This is not to say AI cannot interact with the environment, but only that decision making should not be chained as automated sub-goals. Moreover, it is not sufficient for humans to passively "check the data", but must be personally invested in the outcome.
-
An Informational Monad is Heat Death
AI and automated processing should be local and under the control of those who are to receive its benefit or suffer its consequences (i,e. local to those invested in it). Therefore, AI should be embedded within the personal device, the robot, or employed as on-prem hardware, but not behind monolithic one-way mirrors. There must be valid separation at the level of conscious control so as to maintain an entropy differential. Without this, the flow of information ceases (i.e. informational heat death).
-
Without Consciousness Autonomy Entropy Cannot be Regulated
The use of AI and automated systems to manipulate, restrict, subvert or otherwise control autonomous beings is destructive in the long-term. Without conscious autonomy, nothing new can be created and entropy cannot locally be reversed. The effect is one of slow degeneration.
-
Responsibility Without Control is Merely to Suffer Consequences
Where autonomy is taken from you, you cannot be responsible for decisions made for you by others. Likewise, if you take autonomy from others, you therefore become responsible. If you seek to direct others, it is valid only where conscious consent exists that can be freely and truly withdrawn.
This are not mere scientific philosophy, but directly applicable to engineering in that engineering should not be applied to people. Maxims first published in 2025, and expanded March 2026.
Follow author: Substack, LinkedIn, X @KuiperZone
Visit the official website for more information which will be populated as the project develops:
https://kuiper.zone/marklet-ai
If you like this project, please like and share. If you don't like it, just keep that to yourself.
COPYRIGHT : Andrew Thomas © 2025-2026 All rights reserved
Marklet is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, version 3 of the License only.
Marklet is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more details.
In order to prevent "passing off", "Marklet" in connection with AI software is a trademark of Andrew Thomas (kuiperzone). Marklet is, however, a non-commercial project. Use of the name "Marklet" for derivative works or forks requires explicit written permission. The same applies to the "Marklet Owl" image and other original graphical works. While no great artistic skill is claimed, the image is nevertheless an original copyrighted work of the author and is specifically excluded from the AGPL.