Context Studio
Build a reusable library of files, webpages, YouTube links, and media for conversations and creative work
Context Studio is the new name for Attachment Manager. It is the central place to store, organize, preview, and reuse context across Msty Studio.
Use Context Studio when you want documents, images, videos, webpages, or YouTube transcripts to be available beyond a single chat message.
For the best results, configure Context Studio Settings before building a large context library. In particular, assign metadata models for images and documents so Msty can generate useful summaries, descriptions, and tags for Dynamic Mode and filtering.
Open Context Studio
Open from the sidebar
Select the paperclip icon in the main sidebar to open Context Studio.
Open from a conversation
Select the attachment icon in a chat, then choose Browse Existing Attachments to pick reusable context from Context Studio.
Manage selected context
After Context Studio opens from a conversation, select the items you want to attach and apply the selection back to the chat.
Build Your Context Library
Context Studio supports:
- Documents: upload files and keep reusable text context available across conversations
- Images: store visual references for chats and Media Studio workflows
- Videos: add video files as reusable visual context
- Webpages: save fetched webpage content as reusable context
- YouTube links: import video details and transcript context
You can drag files into Context Studio, upload with Add Attachment, add websites with Add Website, or import YouTube links with Add YouTube.
Organize Context
Use folders to keep large context libraries manageable.
- Create root folders from the Context Studio overflow menu
- Drag files and folders to reorganize them
- Use quick filtering to find folders or files by name
- Sort folders and files by name, creation date, or update date
- Use favorites to keep important items easy to find
Folder metadata can summarize what a folder contains. This improves filtering and helps Dynamic Mode decide which context is relevant.
Preview and Inspect Items
Select an item to open its inspector.
The inspector can show:
- Image, video, webpage, and YouTube previews
- Converted text previews for supported documents, webpages, and YouTube transcripts
- File type, size, added date, and updated date
- Short description, long description, and tags
- Associations showing where the item is used
Metadata fields auto-save when edited. You can also process selected items to generate metadata with the configured metadata models.
Reuse Context in Conversations
There are two ways to use Context Studio content in a chat:
Select Specific Attachments
Use the attachment icon in the chat input, choose a content type, and select the items you want included.
Selected attachments are available to the model for that conversation and can be selected or unselected as needed.
Use Dynamic Mode Aurum Perk
Dynamic Mode lets a tool-capable model search your Context Studio library on demand instead of pushing every selected item into the prompt upfront.
Use Dynamic Mode when:
- The library or folder is too large to include directly
- You want the model to find only the most relevant files
- You have good metadata, summaries, and tags on your context
- You want to scope retrieval to specific folders or the full library
Dynamic Mode is available with Aurum or Enterprise access. It requires a model that supports tools. If the selected model does not support tools, choose specific attachments instead.
Use with Media Studio
Context Studio and Media Studio work together.
- Start a new Media Studio session from supported image or video attachments
- Reuse stored images, videos, and documents as Media Studio context
- Keep generated media available for later conversations when exported with the session
See Media Studio for image and video generation workflows.
Configure Context Studio
Open Context Studio Settings from the Context Studio settings action.
Settings include:
- Metadata on New Uploads: automatically generate metadata for new documents and images when metadata models are assigned
- Metadata Processing Models: assign vision and document metadata models through Model Squad
- Image Reduction Defaults: control default image reduction mode and quality
- Web Fetch: choose Studio's default website fetch or Jina, set website content format, and tune extraction parameters
- Default Prompts: edit Context Studio prompts in Settings > Default Prompts
For website attachments, Studio's default fetch requires a desktop sidecar connection. If sidecar is unavailable, enable Jina in Context Studio settings and add a Jina API key.