For the complete documentation index, see llms.txt. This page is also available as Markdown.
Settings
The Settings section holds two platform-wide admin stores: System Settings (arbitrary, typed configuration values) and Global Secrets (arbitrary name/value secrets available platform-wide).
System Settings
A System Setting is a name/type/value entry — not a fixed list. A small set of known settings (e.g. a LangFuse observability URL) show up as ready-made templates, but admins can also define any custom setting a workspace or the platform needs to read.
Go to AI Admin → Settings → System Settings.
System Settings, empty
Click + Add and fill in:
Type* — String, Number, Bool, or JSON
Setting* — pick one of the predefined settings, or Custom Setting to define your own
Name* — required for a custom setting; letters, numbers, ., _, - only
Value* — the input control matches the chosen Type (text field, number field, toggle, or a JSON editor)
Labels (optional) — free-form tags for grouping related settings
Description (optional)
Add System Setting form, empty
Add System Setting form, filled in
System Setting added
Settings created here are platform-wide (Global scope). A separate Workspace Settings page, available within a workspace's own settings, lets you define a same-named setting scoped to just that workspace — it overrides the global value for that workspace only, the same override pattern used by LLM Mappings and Quotas.
Global Secrets
Global Secrets are name/value pairs available platform-wide — distinct from Provider credentials, which are tied to a specific cloud/tool integration. A Global Secret is just a value your agents can reference, merged in alongside any Workspace-, Project-, or Ticket-level secrets.
Go to AI Admin → Settings → Global Secrets.
Global Secrets, empty
Click + Add and fill in:
Key* — the secret's name; cannot be changed after creation
Value* — the secret value; hidden by default, with a show/hide toggle
Add Secret form, empty
Add Secret form, filled in
Secret added, value masked
Once saved, a secret's value is always masked in the list and API responses — only the last few characters are shown. If a Workspace, Project, or Ticket defines a secret with the same key, that more specific value takes precedence over the Global one.
