Using the portal
How the myt AI Cloud web app fits together — dashboard, projects, Playground, API keys, and documentation.
This guide is for teams using the myt AI Cloud portal (the signed-in web application). If you are integrating from code only, start with Quickstart and Authentication.
Portal vs programmatic API
| Portal (web app) | Programmatic API | |
|---|---|---|
| Who | People in your organisation | Applications and scripts |
| Sign-in | Passwordless email OTP | API key in Authorization: Bearer … |
| Typical tasks | Try models, manage keys, wallet, members | Production inference, automation |
| Base URL | The app URL you were given (e.g. dashboard, settings) | Your deployment’s inference API root (see Quickstart; often includes a /v1 prefix) |
The portal and the inference API work together: you configure orgs, projects, and keys in the app; your services call the API with those keys.
Organisation and project context
After sign-in you work inside an organisation and an active project:
- Organisation — billing, members, wallet, and org-wide settings.
- Project — where API keys and usage are grouped; switch projects from the top bar tenant switcher.
Many screens (Playground, API keys, usage) only show data for the active project. If something looks empty, check that the correct org and project are selected.
Dashboard (Home)
Route: /dashboard
The home page gives a quick health view for the active organisation (usage, spend, active keys) and shortcuts to common tasks. Use it as the starting point after login.
Playground
Route: /playground
The Playground is the no-code way to try models before you ship an integration. Playground use is free of charge — it does not draw from your wallet balance. Per-turn cost shown in the chat thread is an informational estimate only.
- Choose a model from the catalog.
- Optionally set a system prompt and parameters (temperature, max tokens, and related controls).
- Chat in the thread; responses can stream as they arrive.
- Use View code to copy a starter snippet — replace the placeholder API key with a key you created under API keys.
Playground traffic is for evaluation and debugging inside the product. For production workloads, create a dedicated API key, add credit to your wallet, and call the inference API from your own environment (Quickstart).
Model catalog (portal)
Route: /models
Browse models available to your organisation: capabilities, pricing hints, and detail pages with integration examples. The list reflects what your operator has enabled for you. For the machine-readable list over the API, use GET /v1/models.
API keys
Route: /api-keys
Create and manage virtual keys scoped to the active project:
- Create a key, name it, and store the secret immediately (it is not shown again in full).
- Block or delete keys when people leave or credentials may be exposed.
- Open a key’s usage view for request and spend summaries.
Use these keys in your applications (MYT_API_KEY in the quickstart). Do not embed keys in public front-end code.
Settings (owners and admins)
Route: /settings (gear next to your user menu)
Common areas:
| Area | Route | Purpose |
|---|---|---|
| Profile | /settings/profile | Your account |
| Members | /settings/members | Invite users, roles, project access |
| Projects | /settings/projects | Create projects, model access, limits |
| Wallet | /settings/wallet | Balance, add credit, transactions (billing roles) |
| Activity logs | /settings/activity-logs | Security and admin audit trail |
Developers often use Members and Projects; owners additionally manage Wallet and organisation details.
Support and documentation
- Documentation (this site) — integration guides and API reference; open from the sidebar Help section when configured.
- Support — in-product help at
/supportfor access and account issues.
Suggested workflow for new teams
- Sign in and confirm organisation and project in the top bar.
- Open Playground, pick a model, and run a test message.
- Create an API key for the project.
- Follow Quickstart with that key and your inference API base URL.
- Assign members to the project and set wallet / limits with your operator as needed.
Next
- Quickstart — first API call
- Authentication — bearer tokens and key scope
- Models — choosing a model for your workload