Skip to content

Interface Overview#

OptScale AI Chat is the end-user workspace for day-to-day conversations with registered AI providers. The layout splits navigation and history on the left, account and model controls across the top, and the active conversation with the message composer in the center.

Providers, policies, and organization settings are configured in the Admin UI. For platform context, see Architecture Overview.

Layout overview#

The Chat UI has three regions:

Table 1: OptScale AI Chat — main layout regions
Region Purpose
Left sidebar Start new chats, browse history, and switch areas via the icon rail
Top header Choose the active provider (model) and organization
Main area View the thread, use message actions, and compose the next prompt

Icon rail#

A narrow vertical bar on the far left provides quick navigation between major areas (for example Home, Chat, and Settings). Chat is highlighted when you are in the conversation workspace.

New chat and history#

  • NEW CHAT — Starts a blank conversation at the top of the sidebar.
  • Chats — Lists previous conversations grouped by time (for example Last month, March, February, or month/year labels such as 11.2023).

Each history row shows a short title (for example Translate this, Explain this code, fix typescript errors) and a menu for actions such as rename or delete. Select a row to reopen that thread in the main area.

Some rows may show a document icon when the thread includes attachments or imported content.

Import chats#

Before you import

  • Download the vendor export promptly (ZIP or json format).
  • Finish baseline setup in First Steps (organization, provider, allowed providers).
  • Select the target organization in the Chat header.

Import into OptScale AI Chat

  1. Open AI Chat.
  2. Click Import chats control at the bottom of the Chat sidebar Import chats at the bottom of the Chat sidebar.
  3. Upload the export file — the ZIP or json from the vendor email.
  4. Wait for processing to complete. Imported threads appear in the Chats list with titles taken from the source conversation.
  5. Open a thread to review history, then send new messages with a provider from the header selector (subject to allowed providers and organization policies).

Note

Imported messages are read-only history from the source assistant. New replies in OptScale AI use your organization's registered providers and governance rules, not the original vendor subscription.

For export steps per vendor and what to expect after import, see Use Cases — Migrate chat history.

Top header#

Table 2: OptScale AI Chat — top header controls
Control Location Purpose
Provider selector Upper left of the main area Dropdown for the active model (for example llama3:latest). Outbound messages use this provider until you change it.
Connect to external tools Next to the provider/model selectors Opens Connect this model to external tools with Base URL, API key, and Model name for OpenCode, Cursor, and other OpenAI-compatible clients. See Get connection values from Chat.
Organization Upper right Dropdown for the active tenant (for example Sunflower corporation). Scopes providers, policies, and usage to that organization.
Profile Far right User menu and account options (circular avatar).

Connect to external tools control

Administrators configure which providers appear in the dropdown in the Admin UI; Chat does not register new vendors on its own.

Main chat area#

The center panel shows the active thread.

User messages#

Your prompts appear on the right in a light bubble (for example Hello!), with a User label and avatar so user and assistant turns are easy to scan.

Assistant messages#

Model replies appear on the left, labeled with the provider identifier (for example llama3:latest) and the model icon.

Below each assistant message, a row of actions is available:

Table 3: OptScale AI Chat — assistant message actions
Action Purpose
Speaker Text-to-speech for the response
Copy Copy message text to the clipboard
Edit Edit the assistant message in place
Thumbs up / down Submit feedback on the response
Refresh Regenerate the reply

Exact actions may vary slightly by provider or organization policy.

Message composer#

A wide input field at the bottom of the screen is scoped to the selected provider (placeholder such as Message llama3:latest).

Table 4: OptScale AI Chat — message composer controls
Control Position Purpose
Attach file Left of the field (paperclip) Add file attachments to the prompt
Web search Left of the field (globe) Enable web access for the request when allowed
Send Right of the field (arrow) Submit the message

An additional control on the right (for example a chevron) may expose extra input options depending on your deployment.

Supported actions#

Beyond plain text chat, OptScale AI Chat supports workflows that depend on enabled providers and organization settings:

  • Text conversations — Multi-turn dialogue with the selected provider.
  • File attachments — Upload supported file types with the paperclip control.
  • Web access — Ground or augment prompts with web search when the globe control is available.
  • Voice and audio — Speech-related input and playback where enabled (including listen on assistant messages).
  • Image generation — Create images when the active provider and policy allow it.
  • Audio handling — Process or generate audio content on supported models.

Use NEW CHAT and the Chats list in the left sidebar to start threads, reopen history, and manage conversations. Switch organization or provider from the top header when your role and policies allow it.

Topic Where to read more
Import chats (UI) Import chats (this page)
Export and migrate from ChatGPT, Claude, or Gemini Use Cases — Migrate chat history
Initial setup (organization, provider, Chat) First Steps
Admin UI (providers, policies, FinOps) Core Services
Model Training (tasks, datasets, environments) Model Training, Experiment Tracking, Environments & Operations
Platform architecture and request flow Architecture Overview
Connect OpenCode, Cursor, and other tools External Tools — Get connection values from Chat