
Freddai: Making My AI Workspace Easier to Follow
Switching AI tools had become its own little project. I would change models, find the project again, remember another interface, then deal with the accounts. Logging out on the desktop could mean logging in through the browser and reconnecting my phone too.
That is the friction behind this Freddai walkthrough. I wanted the experience of moving between my tools to feel clearer.
Start with the work in front of me
In the recording, I tour the dashboard and start a thread using voice-to-text. The test request is deliberately small. I want to see the message arrive and understand what happened next.
There is a playful animation of the agents at work too. It is still rough in the recording. I enjoy building that part, but the useful question underneath it is simple: can I see whether something is happening?
I also show a daily recap. I point out that it is incomplete because I am still using other apps alongside Freddai. That limit matters when reading a summary of the day.
Find the conversation again
The thread-manager part of the video starts from another small frustration: project names and conversation names can be difficult to distinguish.
I show project headings and thread status, then move between active and recent conversations. I want to be able to find what needs attention without scrolling through everything I did earlier.
The same question applies to usage information. Before starting another request, I want to understand which account I am using and when its usage resets. In the demo, I show that information and the process of choosing another subscription.
This is an account-management workflow in the recording. It does not make a provider's usage limits disappear.
Make the next request explicit
The new-thread picker is where I choose the project, agent and model. Then I can write or dictate the request.
I also demonstrate the rocket shortcut. It starts a thread asking the agent to commit and push, with lint and tests passing first. I treat that as a request to the agent. The shortcut itself does not prove that the checks passed or the push succeeded.
That distinction is useful to keep in view whenever a convenient button starts work with consequences. I still need to understand the result.
Keep the unfinished parts visible
The published description calls Freddai a work in progress. Mac, Windows, Linux, iPhone and iPad are the plan, and the mobile apps still need Apple review. Those are the availability qualifications attached to the video, rather than a claim that every version is ready today.
What interests me in this experiment is the everyday friction around working with agents. Where did the conversation go? Is it still running? Which account am I about to use? The walkthrough shows the interface decisions I was trying in response.
