Organisation settings
Path: Settings → Organisation (URL /app/org-settings).
A grab-bag of org-wide configuration. Sub-pages within:
Overview
The org's basic info: name, logo, primary contact email, default language, default timezone.
Branding (Full tier)
White-label settings:
| Field | Notes |
|---|---|
| Org name | Replaces "Bee Flow" in titles. |
| Logo | Replaces the bee mark. SVG / PNG up to 1 MB. |
| Primary colour | Hex. Replaces amber as the accent. |
| Secondary colour | Hex. |
| Custom domain | E.g. ai.example.com. DNS CNAME to beeflow.ai. |
| Custom email sender | From: noreply@example.com for invites / notifications. |
White-label is gated by the white_label feature flag — Full tier only.
Defaults
What new agents / KBs / users inherit:
| Default | Purpose |
|---|---|
| Default model | What new agents use unless overridden. |
| Default Privacy Shield level | What new agents inherit. |
| Default integrations enabled | Mirrors the NC integrations panel. |
| Default starter prompts language | English / Dutch / German / etc. |
Studio policy
Org-wide rules for what users can do in Studio:
| Rule | Notes |
|---|---|
| Allow user agent creation | Off → users can use existing agents but not create new ones. |
| Mandate templates | New agents must start from a template. |
| Require admin approval for publish | Public-marketplace submissions need admin sign-off. |
| Lock per-user integration overrides | Users can't re-enable disabled tools. |
Licence & usage
Path: Org settings → Licence & usage.
| Field | Notes |
|---|---|
| Active tier | Read-only — derived from the licence. |
| Licence key | Paste / clear here (details). |
| Tier limits | Users / agents / msgs per month. |
| Current usage | This calendar month. Resets at 00:00 UTC on the 1st. |
| Last verified | When the server last polled the licence-server. |
| Refresh now | Manual re-check button. |
Privacy
Org-level Privacy Shield settings (the same JSONB record described in Privacy Shield).
| Field | Notes |
|---|---|
| Master toggle | Turn the shield off entirely (not recommended). |
| Detection backend | Azure / Local / Both. |
| Confidence threshold | 0–1, default 0.7. |
| Custom regex categories | Org-defined patterns. |
| EU mode | Minimises logs, anonymises IPs. |
| Lock user override | Users can't tune their own privacy level. |
DLP (Enterprise+)
Configure policy rules for prompt + tool-result scanning. See DLP & guardrails for the action types and audit log structure.
SSO (Enterprise+)
| Provider | Setup |
|---|---|
| SAML 2.0 | Paste IdP metadata XML; Bee Flow exposes ACS URL + entity ID. |
| OIDC | Set issuer + client ID + secret. |
When SSO is enabled, the username/password form is hidden by default. Set ALLOW_PASSWORD_LOGIN=true to keep it as break-glass.
Audit (Enterprise+)
| Field | Notes |
|---|---|
| Retention days | How long to keep guardrail_events. Default 90. |
| SIEM webhook | URL + shared secret for real-time event push. |
| Email alerts | Email address(es) for high-severity violations. |
See Audit & compliance for the data model.
Voice (Pro+)
| Field | Notes |
|---|---|
| STT provider | Voxtral / Deepgram / Mistral / Whisper local. |
| TTS provider | Voxtral / ElevenLabs. |
| Default voice | Picker from your provider's voices. |
| Allow voice calls | Master toggle. |
Integrations (other than NC)
A panel listing every integration the server is configured for (env vars set), with org-level toggles + per-group overrides — same model as the NC integrations panel.
Memory
| Field | Notes |
|---|---|
| User memory enabled | Master toggle for the Memory Extractor system agent. |
| Org memory | Shared memory facts visible to all agents (e.g. "We refer to Q1 as 'the launch quarter'."). |
| Retention | Forever / 12 months / 6 months. |
Danger zone
| Action | Effect |
|---|---|
| Reset wizard | Re-runs the first-run wizard on next login. |
| Delete all conversations | Org-wide chat purge. Irreversible. |
| Delete organisation | Tombstones the org. 30-day grace period before hard delete. |