What Stuga sends and stores
For the person who set up Stuga.
For the person who set up Stuga. Stuga sends no telemetry. Requests come from checking releases, offering samples and features you use.
Check requests made without setup
Section titled “Check requests made without setup”Stuga asks GitHub for its release list daily. The request carries no installed release number, Stuga ID or workspace content. Turn off Check for new versions in Settings → This node → About to stop it.
Opening Create a workspace fetches the sample list, cached for up to an hour. Choosing a sample downloads it. These come from GitHub unless you configure another source.
Check what your features send
Section titled “Check what your features send”| Feature | What leaves Stuga’s computer |
|---|---|
| Built-in AI and Reranking | The chosen service receives questions or conversations, applicable instructions and content used to answer, such as documents, passages, rows or chat images. |
| Semantic search | The chosen service receives text from documents not hidden from search, when set up and as text changes, plus searches and questions. Changing its AI prepares the documents again. |
| Notifications | Your selected chat service, webhook (an address that receives events) or email server receives the title, text and link; webhooks and email also receive the recipient. |
| Workspace webhooks | The configured address receives workspace events. |
| Your organization’s sign-in service | Stuga contacts the service during setup and sign-in. |
| Images and files added by web address | Stuga fetches them from that address, stores them and serves the copies. |
| Connected apps | An app may supply a web address Stuga reads to display its name when you approve it. |
An outside AI tool, such as Claude Desktop, sees the Stuga content it reads and writes.
Test and the Model list contact the AI service without workspace content.
For built-in features, text stays on Stuga’s computer when all AI services used run there. Hide from search in a document’s ⋯ menu keeps its text out of semantic search; an AI tool can still read it if given access.
Installing and updating download Stuga from GitHub. Docker also downloads images from GitHub’s container registry.
Check what Stuga stores
Section titled “Check what Stuga stores”Documents, databases, files, sharing, history, accounts and service keys live on Stuga’s computer. Anyone who can read its disk can read the document data and stored service keys. Backups contain these too. Account passwords are stored as hashes, scrambled copies used to check sign-in.
In Settings → This node → Storage, choose how long activity history, AI usage records and idle Ask threads are kept. Trashed documents and databases are deleted after 30 days. Their images can remain on disk until reclaimed.
- Set up built-in AI and semantic search — choose your services.
- Back up and restore — keep another copy of your data.
More detail for people who run servers: What Stuga sends and stores on GitHub.