Use the admin console
The admin console is where you manage Orchestrate. Everything it shows comes from the API, so anything you can do here you can also script. See the API reference.
Sign in
Open the console address your operator gave you and choose Sign in. The browser goes to the sign-in page, where you can:
- enter your username and password, or
- choose your organization's identity provider, if one is connected. See Identity and single sign-on.
If two-factor sign-in is required for you, you set up an authenticator app the first time and enter a code each time after. If you were invited with a link, open it to choose your password first.
There is no self-service sign-up: an administrator invites you or creates your account and gives you a role. See Users, roles and projects.
Your name and your highest role (viewer, developer or admin) show at the bottom of the sidebar. Choose Sign out there when you are done.
Pick a project
Most pages show one project at a time. Choose the organization and project in the picker at the top of the sidebar. The console remembers your choice in this browser.
Every install has a project called Default.
Language
Choose English, Arabic or French from the menu in the top bar, or under Settings > Language and region, where you also set your time zone. Arabic is laid out right to left. See Languages.
Pages
The sidebar groups pages by what you do.
Build
| Page | What you do there |
|---|---|
| Workflows | Build multi-step workflows on a graph canvas. See Workflow graphs. |
| Agents | Create single-step agents and code agents, see their versions, choose which is live, and connect them to people and applications. See Agents and Code agents. |
| Playground | Chat or talk with a model or an agent, inspect each turn and save a turn as an eval case. See Playground. |
| Knowledge | Create knowledge bases, upload documents and test retrieval. See Knowledge bases. |
| Connectors | Connect databases, APIs and MCP servers so agents can use them as tools. See Tools and connectors. |
| Voices | See the voices the project's agents can speak with, preview them, and clone a voice. See Voices and speech models. |
| Connected accounts | Connect your own accounts for connectors that act as each user. See Act as each user. |
Run
| Page | What you do there |
|---|---|
| Runs & traces | Follow runs and open one to see its steps, tool calls, approvals and feedback. See Runs and traces. |
| Models | Browse the model catalog, add models from the library, a URL or your own container, deploy them and restrict them to projects. External providers connects hosted providers. See Models and serving and External providers. |
| Realtime | See voice sessions and phone calls and their event logs. See Voice agents. |
| Telephony | Manage SIP trunks and phone numbers, place calls, and see calls and recordings. See Phone calls (SIP). |
| Clusters | Connect other Kubernetes clusters that models can be built and deployed on. See Models and serving. |
| Sites | Manage sites on customer premises in hybrid mode. See Hybrid sites. |
Govern
| Page | What you do there |
|---|---|
| Governance | See blocked actions, pending approvals, the rules in force and custom policies, and open Compliance and the AI registry. See Policy and Compliance. |
| Evals | Write eval suites, run them, compare runs, and follow online evals and feedback. See Evals and Online evals and feedback. |
| Usage | See model calls, tokens and spend by model, project, agent, key and day, and set model prices. See Usage controls. |
| Audit log | Read and verify the audit log, export it to a SIEM and set its retention. See Audit log. |
Admin
| Page | What you do there |
|---|---|
| Organizations | Manage organization members, settings (time zone, default guardrail profile, model policy, branding), MFA and budget. |
| Platform | Onboard and offboard organizations, on an install that hosts several. Platform operators only. See Multi-tenancy. |
| Projects | Create folders and projects, manage members, budgets and limits, model access and guardrail profiles, and delete or restore projects. See Users, roles and projects. |
| Identity & SSO | Connect identity providers, manage groups and their roles, require two-factor sign-in, and invite people. See Identity and single sign-on. |
| API keys | Create, rotate and revoke your personal keys for the Model Gateway, and agent keys. See Call the Model Gateway. |
| Licence | See and upload your licence, its features, limits and usage, and set code-agent quotas per project. See Licence. |
| Settings | Your own language and time zone. Platform admins also set the install's default language, external providers, voice capacity, branding, voice catalog and service keys. See Settings and branding. |
| System | The version, the health of every component, the licence, vulnerability scanning, and the facts an offline upgrade needs. Read-only, for platform admins. See Operations. |
The Overview page shows active and recent runs, model calls per day and the most used models, all from your install.
Settings and branding
On Settings, a platform admin can:
- set the console's default language for the install
- switch external providers off and on again, within what the Helm values allow. See External providers.
- set the branding: the product name (up to 40 characters), a logo for light and dark themes (PNG, SVG, JPEG or WebP, up to 256 KB) and a primary colour. The console, the sign-in page, page titles and hosted chats use it.
- add voices and speech models to the voice catalog
An organization admin can override the branding for their organization under Organizations > Settings.
What you see depends on your role
The console hides or disables actions your role cannot take, and says why. For example, deploying a model needs the developer or admin role, and creating an agent needs the developer or admin role on that project. If a page says you do not have access, ask an administrator for the role you need.