Start with one task
Make the tool you keep working around.
Describe a useful small app and try a separate preview. The quote calculator example on the homepage shows how materials, hours and VAT can come together in a tool for your own work.
Product roadmap · October 2026
Try the public Linux alpha today. The connected workspace is available now; optional local goal runs are available as an experimental feature.
The product today
Keep each app close to the conversation and decisions that shaped it.
Build a small browser tool from a task you know.
Overview, Conversations and App in one project.
Turn them on in Settings to give a local model a goal and a few checks; inspect the result because this opt-in feature is experimental.
Explore longer goals and contained app changes.
Your project
Each project keeps its brief and direction in Overview, the discussion in Conversations, and building in App. Return later with earlier decisions close by.
Start with the public alpha →Start with one task
Describe a useful small app and try a separate preview. The quote calculator example on the homepage shows how materials, hours and VAT can come together in a tool for your own work.
One project, three connected spaces
The workspace keeps the project brief and direction in Overview, the discussion in Conversations, and app-building in its own focused space. Previous work stays visible so you can pick up the next step with context.
A local model, working toward your goal
Set a goal and checks that fit the task, such as creating a file or including required JSON keys. The model can work through small changes, check its progress, and return control when it reaches a limit or needs your input.
Early real-model testing is under way. We will use what we learn to shape longer goals and more involved app work.
Build on what works
As local workflows prove useful, we will look at longer tasks and a broader set of app changes—keeping progress visible and making it easy to review the result.
Use the runtime that fits
Fictor supports local Ollama models today, with optional cloud runtimes for tasks where they fit. Your runtime choice stays part of a project designed to remember the work around each result.
Read about security and runtime choices →