Configuration Overview
All of Silicon Worker's configurable options live under Settings in the app. This page maps what you can configure by area, what it affects, and links out to the matching guide or reference page so you can find things quickly.
This is a navigational reference. The exact fields depend on the options your current version's Settings screen actually shows; open the links below for the detailed steps in each area.
Settings at a glance
| Settings area | What you can configure | Related pages |
|---|---|---|
| Models / Providers | Add a provider (choose the protocol type); fill in Base URL, API Key, and model name; pick a primary model plus one or more backup models for failover. | Model Providers · Configure providers · Connect Your First Model |
| Permission mode & risk confirmation | Set how much Silicon Worker asks for your consent first (Free / Controlled / Restricted), controlling the confirmation behavior before sensitive actions. | Permission modes · Tools & risk · Permissions |
| System permission grants | Grant OS-level permissions on demand (accessibility, automation, full disk access, calendars, reminders, notifications, and so on) to enable desktop/app capabilities. | System Permissions · Grant system permissions |
| Scheduled tasks | Create one-off or recurring scheduled tasks, with a trigger schedule and the task description to run. | Scheduled tasks & remote |
| Desktop & browser automation | The automation capability itself works after install; the system permissions it needs are granted under System permissions. | Desktop & browser automation · System Permissions |
| Memory | Maintain long-term memory: review, add to, or clean up private memory and global memory. | Memory · Manage memory |
| Knowledge base | Create knowledge bases, import material, and bind them to a session, expert, or team for retrieval. | Knowledge base |
| Skills | Install, view, and manage skills (standard procedures in SKILL.md form). | Skills · Using skills |
| Plugins | Install a plugin to bring in its packaged skills, experts, hooks, and MCP servers in one go. | Using plugins · Using skills |
| MCP | Add MCP connectors to bring in external tools: by connection type (stdio / HTTP / SSE), fill in the launch command, address, environment variables, or headers; services that require authorization can use OAuth. | Connect MCP |
| Chat-tool connection (remote) | Connect a chat tool to send tasks and receive results remotely from your phone or another device. Available connectors depend on your version. | Scheduled tasks & remote |
Models & providers
Under Settings → Models / Providers, you can add one or more providers and configure models:
- Provider protocol type: OpenAI-compatible (the default, covering nearly all services) or Anthropic native.
- Base URL, API Key, model name: connecting a model usually only needs these.
- Primary & backup models: pick a primary model for the session, then set one or more backups; when the primary hits a transient problem such as a network error, rate limit, or timeout, it retries automatically and, if it still fails, switches to a backup.
See Model Providers and Configure providers.
Permissions & safety
- Permission mode: set the confirmation level across Free, Controlled, and Restricted (see Permission modes).
- System permissions: grant accessibility, automation, full disk access, calendars, reminders, notifications, and so on as needed (see System Permissions).
TIP
A permission only means "this capability is allowed." Whether a given high-risk action actually runs is still gated by the tool's risk level together with your permission mode. See Tools & risk.
Automation & scheduling
- Scheduled tasks: have a task fire automatically on a schedule, one-off or recurring. Scheduled tasks run on your machine, so the app must stay running and the device powered on and online.
- Desktop / browser automation: the capability works after install; the system permissions it needs are granted under System permissions.
See Scheduled tasks & remote and Desktop & browser automation.
Memory & knowledge base
- Memory: long-term memory is split into private (owned by an expert) and global (shared across roles); you can add to or clean it up (see Manage memory).
- Knowledge base: create, import, and bind to a session, expert, or team for retrieval (see Knowledge base).
Skills & plugins
- Skills: install, view, and manage standard procedures in SKILL.md form (see Using skills).
- Plugins: installing one brings in its packaged skills, experts, hooks, and MCP servers in one go (see Using plugins).