Settings File¶
Zwischen reads all configuration from ~/.zwischen/settings.yaml. This file is created by the install wizard and can be edited directly at any time. Settings are re-read on every session start, so changes take effect immediately without restarting the daemon.
Structure¶
The settings file has four top-level blocks:
env:
ZWISCHEN_API_KEY: "your-key-here"
projects:
- path: ~/work/sensitive-repo
enabled: false
- path: ~/work
enabled: true
defaults:
model: gpt-5.4
api_key: your-provider-key
profiles:
code-review:
agents: [coder, web-dev]
model: claude-sonnet-4-6
env¶
Environment variables injected at startup. ZWISCHEN_API_KEY is required. Values here are defaults — real environment variables take precedence.
projects¶
Per-directory enable/disable rules. See Project Rules.
defaults¶
Baseline configuration for all reviewed agents. See the sections below for all available keys.
profiles¶
Named configuration overrides for specific agents. See Profiles.
Settings resolution¶
Settings resolve in order, with each layer deep-merging on top of the previous:
- Server defaults — built into the Zwischen server
defaults— your baseline configuration- First matching profile — overrides for specific agents
For nested dicts like cycle and tools, the merge is key-by-key. A profile that sets tools.output_visible doesn't erase your defaults.tools.pre_review — both are preserved.
Exception: system_prompt and system_prompt_mode do not deep-merge. A profile's system_prompt is used as-is — it does not inherit from or append to defaults.system_prompt.
Minimal example¶
The smallest working configuration:
This reviews all agents with the default server profile, using GPT-5.4 as the reviewer model.
Next steps¶
- Project Rules — enable or disable Zwischen per directory
- Profiles — configure different reviewers for different agents
- Tool Visibility — control what the reviewer can see
- Models and Auth — model selection and authentication
- Full settings reference — every key in one table