Refresh playground docs for threads and files

docs/toolhive/guides-ui/playground.mdx

@@ -36,6 +36,19 @@ MCP servers directly through natural language commands. You can list servers,
 check their status, start or stop them, and perform other management tasks using
 conversational AI.
 
+### Threaded conversations
+
+Keep separate testing sessions in their own chat threads. The sidebar groups
+your conversations into **Starred** at the top and **Recents** below, so the
+threads you use most stay within reach. Use **New chat** to start fresh, give
+each thread a descriptive name, and remove old runs when you're done.
+
+### Attachments
+
+Send images and PDFs along with your prompts so the model can analyze
+screenshots, diagrams, logs, or reference documents while it calls your MCP
+tools. Each message accepts up to 5 files, and each file can be up to 10 MB.
+
 ## Getting started
 
 To start using the playground:
@@ -93,6 +106,23 @@ To start using the playground:
    have access to all enabled MCP tools and can execute them based on your
    requests.
 
+5. **Manage chat threads**: Use the sidebar to keep separate threads for
+   different testing scenarios.
+   - Click **New chat** at the top of the sidebar to start a fresh thread.
+   - Double-click a thread row, or open its **Thread options** menu and choose
+     **Rename**, to give it a descriptive name. New threads show as `New chat`
+     until you rename them.
+   - Choose **Star** in the **Thread options** menu to pin a thread under
+     **Starred** at the top of the sidebar; choose **Unstar** to remove it from
+     the starred list.
+   - Choose **Delete** to remove a thread. The playground asks
+     `Delete "<THREAD_NAME>"? This cannot be undone.` before removing it.
+
+6. **Attach images or PDFs**: To send a file with your message, drag it onto the
+   playground or open the composer toolbar menu and choose **Add images or
+   PDFs**. You can attach up to 5 files per message, and each file must be 10 MB
+   or smaller. Supported types are images and PDFs.
+
 ## Using the playground
 
 ### Testing MCP server functionality
@@ -173,6 +203,64 @@ The playground shows detailed information about each tool execution:
 This visibility helps you understand exactly how your MCP servers are behaving
 and identify any issues with tool implementation or configuration.
 
+### Manage playground threads
+
+The playground keeps each conversation in a separate thread so you can run
+several testing sessions in parallel without losing context. Open the sidebar to
+see your threads, with **Starred** entries pinned at the top and **Recents**
+below. Untitled threads show as `New chat` until you give them a name. Each row
+shows a relative timestamp (`just now`, `Nm ago`, `Nh ago`, `Nd ago`, or a short
+date for older threads) so you can spot recent activity at a glance.
+
+To work with threads:
+
+- **Start a new thread**: Click **New chat** at the top of the sidebar.
+- **Rename a thread**: Double-click the thread row (the tooltip
+  `Double-click to rename` confirms the action), or open its **Thread options**
+  menu and choose **Rename**. You can also click the title or the pencil icon at
+  the top of the chat to rename the active thread.
+- **Star or unstar a thread**: Click the star icon next to the thread title, or
+  open **Thread options** and choose **Star** or **Unstar**. Starred threads
+  appear under **Starred** at the top of the sidebar.
+- **Delete a thread**: Open **Thread options** and choose **Delete**. The
+  playground asks `Delete "<THREAD_NAME>"? This cannot be undone.` before
+  removing it. Confirm with **Delete**, or back out with **Cancel**.
+
+### Attach files to a message
+
+Add images and PDFs to a message so the model can read them while it works with
+your MCP tools. The composer accepts up to 5 files per message, each 10 MB or
+smaller, and supports image files and PDFs.
+
+To attach files:
+
+1. Open the composer toolbar menu and choose **Add images or PDFs**, or drag
+   files onto the playground window. Drag-and-drop is enabled across the entire
+   playground.
+2. Type your prompt and send the message. If you send a message that only
+   contains attachments, the playground records the message text as
+   `Sent with attachments`.
+
+In the chat history, the playground previews each attachment alongside the
+message:
+
+- **Images** appear inline. Click an image to open it in a larger modal preview.
+- **PDFs and other non-image files** show as `📎 <FILE_NAME>` with a
+  **Download** link so you can save the original file.
+
+If a file is rejected, the playground shows a toast that explains why:
+
+- `You reached the maximum number of files` /
+  `You can only upload up to 5 files` when you exceed the per-message limit.
+- `File size too large` / `The file size must be less than 10MB` when a file is
+  over 10 MB.
+- `File type not supported` / `Only images and PDFs are supported` when the file
+  isn't an image or a PDF.
+
+The composer placeholder reflects the playground state. It shows
+`Select an AI model to get started` when no model is selected, then
+`Type your message...` once you've chosen one.
+
 ## Recommended practices
 
 ### Provider security
@@ -201,6 +289,16 @@ and identify any issues with tool implementation or configuration.
 4. **Error handling**: Intentionally trigger error conditions to validate proper
    error handling
 
+### Thread and attachment hygiene
+
+- Delete unused threads so the sidebar stays focused on the work you actually
+  return to.
+- Star the conversations you want to keep close at hand. Otherwise they get
+  pushed down as new chats arrive in **Recents**.
+- Treat attachments as inputs that may leak data. Strip credentials, customer
+  information, and other sensitive content from PDFs and screenshots before
+  sharing them with an AI provider.
+
 ## Next steps
 
 - Learn about [client configuration](./client-configuration.mdx) to connect