Where Does My Research Data Go?
Scholardo stores your project on your computer and does not keep a cloud copy. Whether task content leaves the machine depends on the services you choose.
| Action | Data flow | Local alternative or control |
|---|---|---|
| Read saved files, annotate, render citations | Local project files and indexes | No AI key required |
| Ask a cloud model or use cloud editing | Task text goes directly to your configured provider | Choose a local model |
| Index with cloud embeddings | Document chunks go to your embedding provider | Choose local embeddings; remote indexing needs consent |
| Use Ollama for generation and embeddings | Processing on your machine | Check both configurations separately |
| Fetch papers, RSS or web pages | Requests reach those sources | Read saved copies offline |
| Activate and refresh a license | License service receives activation information | See the license's offline period |
Local-first is not a cloud-model privacy guarantee
A project can be stored locally while its passages are sent to a cloud model for a task. Review your provider's retention and training terms and your institution's requirements before processing confidential or unpublished data. Local models also need appropriate access controls on your computer.
Project tools and commands have different permissions
Scholardo's project-file tools are scoped to the project and approved external sources. Commands and third-party CLIs use their own permissions under your account. Do not treat the MCP path check as an operating-system sandbox.
License and trial
Starting a trial requires an email verification code, not a product account or payment details. Activation requires a connection; fully offline an activated app works for at least two weeks before becoming read-only until it connects again. See License Activation.
The app sends no usage telemetry. The website has its own page-view measurement; website analytics and research-file handling are different things. The authoritative details are in the Privacy Policy.
Before a sensitive project
Choose a local generation model and local embeddings if text must stay on your machine. Review external sources and any CLI permissions. Confirm your institution's approval requirements; neither local storage nor a successful output check certifies compliance.