Effort and output styles
How hard SoreQen thinks before it acts, and how it writes to you while it works.
Effort
| Level | What it means |
|---|---|
| Low | Quick and light. For small, clear changes. |
| Medium | Balanced. The everyday setting. |
| High | Plans first and checks each step. |
| Extra | Investigates deeply. For tricky work. |
| Max | Exhaustive, and reviews its own changes. |
| Apex | The most thorough. For work that must be right. In the desktop app. |
Medium is the default. Change it with /effort, or with --effort (or -e) when you start:
/effort high
soreqen -e extra
Each level is matched to the thinking levels of the model you picked, and the effort menu in the desktop app says which. Thinking is never switched off.
Apex
The top level, in the desktop app's effort menu. Apex prefers the more robust approach even when it is slower, proves each claim with a check rather than an argument, and before finishing reviews its own work as a demanding reviewer would and fixes what that review would reject. The command-line tool offers Low to Max.
Output styles
A style changes how SoreQen writes to you, not what it does.
| Style | How it writes |
|---|---|
default | Brief progress notes and a structured report at the end. |
proactive | Takes initiative on the obvious follow-through, and says what it did. |
concise | As few words as possible. Short updates and a two-line summary. |
explanatory | Explains the reasoning behind choices and the patterns in the codebase. |
learning | Teaches as it goes and leaves small, meaningful pieces for you to write. |
/style concise