> ## Documentation Index > Fetch the complete documentation index at: https://agents.concierca.ch/llms.txt > Use this file to discover all available pages before exploring further. # Settings > Personal and company settings: account, preferences, users, departments, branding, billing, and monitoring. The Settings panel manages everything about you and — for admins — your company. **Where:** Sidebar › **Settings**. The entry opens a panel with grouped sections (⌘K also toggles it; Esc closes). Sections your role can't open are shown greyed with a padlock instead of hidden. ## Section map | Group | Section | Who | | ------------------------------- | ----------------------------------------------------------------------------------------- | ------------------------------------------------------------ | | **Personal** | Overview, Account | Everyone | | **Team** | Users, Departments, Permissions, Knowledge/Memory/Skill Categories, Preferences, Branding | Company Admin | | **Security** | Ingest Keys *(admin)*, API Keys, MCP Access | See rows below | | **Data** | Data Sources *(admin)*, Retention | Admin | | **Integrations** | Telegram, Connections | Everyone | | **Finance** | Billing | Company Admin | | **Provider / Platform Cockpit** | Agency- and platform-level management | Agency Admin / Superadmin — outside the scope of this manual | *** ## Personal ### Overview Your profile card (name, email, role) plus quick-access tiles to the most-used sections. ### Account * **Profile** — display name (saved to your profile), plus an avatar icon and color picker *(stored per device)*. Your login email is shown read-only. * **Timezone** — the timezone your scheduled agent runs fire in; defaults to your browser's. * **Change Password** — new password (min 8 characters) with confirmation. ## Security ### API Keys (the Vault) — everyone Your **personal** provider keys — OpenRouter, Anthropic, OpenAI, Google AI, Groq, Telegram Bot, Slack, Unipile, Apify, Composio, Resend, Stripe, or Custom HTTP — all stored encrypted and never shown in full after saving. * **Add API Key:** provider, auth type (API Key / Bearer Token / OAuth Session / Basic Auth), label, value, optional metadata. * Per key: rename, edit metadata, delete (*"Any agent using this key will lose access immediately."*). * Keys shared company-wide by an admin carry a **Shared** badge and are read-only for members. * **Setup Telegram Bot** — a guided 3-step wizard: paste your bot token, send the bot a message so it detects your chat, done. ### MCP Access — everyone Connect external MCP clients (Claude Code, claude.ai, and others) to Concierca AI **as you** — every call is scoped to what you can see. * **Generate New Key** — label it, then toggle exactly which tools the key may use, and optionally per-table **write access**. The plain key is shown **exactly once**; a ready-made one-line Claude Code command is provided. * Key list shows label, prefix, tool count, last use, and badges (**Connector** for OAuth-created keys, **Stale** for keys unused over 30 days). You can edit a key's enabled tools anytime, or **Revoke** it. * **Connect claude.ai** — no key needed: paste the provided connector URL into claude.ai's custom-connector settings and sign in with your Concierca AI login. ### Ingest Keys — Company Admin Bearer tokens for pushing data into your workspace via HTTP (`POST /api/ingest/`). Generate a key with a label and per-endpoint scopes (or a master `*` scope), copy it once, revoke anytime. Revoked keys stay listed for audit. ## Integrations ### Telegram — everyone Your personal Telegram link: webhook status (with re-register), your chat ID (auto-**Detect** or manual entry with guidance), **Send Test Message**, and **Unlink**. Once set, new agents default to your chat for Telegram delivery. ### Connections See the [Connections](/guide/connections) chapter. ## Team *(Company Admin)* ### Users Manage team members: the table shows Name, Email, Position, Role (Company-Admin / Member), and last activity. * **Invite User** — full name, email, optional position, role (**Admin** — full control / **Member** — standard / **Viewer** — read-only). The system creates the account (depending on setup, with a one-time temporary password to share). * Edit a user's name, email, position, and role; **Remove** deactivates the account (*"The user will lose all access."*). ### Departments Your org structure. Each department has a slug, name, icon & color, and a **visibility** mode: **Open** (all members), **Restricted** (assigned members only), or **Private** (owner + explicit admins). * **Members** dialog — add/remove users per department with a per-department role (admin / member / viewer). * **Visible Data Tabs** — per-department switches controlling which data views the department's members see (views without any assignment stay visible to everyone). * Deleting a department leaves its knowledge/memories/skills unassigned (cannot be undone). ### Permissions A read-only access matrix showing what Admin / Member / Viewer can do per resource type. *Editable per-role overrides are planned but not yet available.* ### Knowledge / Memory / Skill Categories Manage the category lists used across knowledge, memories, and skills: name, description, icon & color, drag-order. Skill categories additionally carry a scope (Global / Department / User). ### Preferences Company-wide AI defaults: | Setting | Meaning | | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Default Agent Model** | Inherited by every agent without an explicit model (default: Auto) | | **Self-Learning (Company Brain)** | One switch that provisions a nightly learning sweep plus a knowledge summarizer per active department; disabling turns them off without deleting anything. Shows the estimated weekly model budget. | | **Max Output Tokens** | Default output budget (1,000–32,000) | | **Budget Warning Threshold** | When to warn about monthly spend (50–100%) | ### Branding Your workspace's look: app name, legal entity, support email, primary/accent colors, logo & favicon URLs, the **AI Persona** (system-prompt prefix for chat), and the chat page title/subtitle — with a live preview. The custom domain is configured by the platform team. Companies managed by an agency see their inherited branding read-only. ## Data *(admin)* ### Data Sources Your standard catalog tables (always active) and any ingest-enabled views, each with its **ingest endpoint** (`POST /api/ingest/`), the request body schema, and per-table API key management — so external tools (e.g. Make.com) can push data in. Rows the ingest rejects (schema mismatch, missing identity, wrong scope) are never silently dropped — they land in **Failed Imports**, a dead-letter list you can review (also readable via the API's `failed_import_list` tool) to see exactly what was rejected and why. ### Retention Company-wide data retention: set *Days to keep* and a nightly job permanently deletes mission runs, agent runs (including prompts), and artifacts older than the window. Leave empty to keep everything forever. Enabling it requires an explicit confirmation — deleted data cannot be recovered. *(This section is currently reachable for platform admins only.)* ## Finance *(Company Admin)* ### Billing Everything about credits and spend, on one page: | Card | What it shows | | ------------------------ | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Plan** | Available monthly plans with included credits, agent and department limits; upgrades apply immediately, downgrades at period end | | **Buy Credits** | One-time top-up packages — credits never expire | | **Module Subscriptions** | Add-on module access with a **Manage Subscription** portal | | **Seat Subscription** | Monthly per-active-user pricing, billed separately from credits | | **Credit Balance** | The fuel gauge: remaining credits and plan rate. Credits are charged per billable agent run (schedule / webhook / email / channel chat), scaled by what the run does — **in-app chat is free** | | **Credit Usage** | The per-run consumption log — click a row for the underlying run | | **Monthly Budget** | A hard USD cap on model spend; runs are blocked at 100%, with alerts at 70 / 85 / 95 / 100% | | **Current Month Usage** | Spend so far (resets on the 1st) | | **Alert History** | Which budget alerts fired this month | The sidebar's credits card mirrors the Credit Balance at all times.