FAQ

Questions people ask before exporting and backing up AI work history.

Does DataMoat upload my AI work traces?

No. The memory, search flow, state, supported skills, supported attachments, and memory encryption keys stay local. The website may call GitHub Releases for download metadata, but DataMoat is not a transcript cloud service.

Which sources are supported?

DataMoat supports Claude CLI, Claude Desktop local-agent sessions, Codex CLI, Codex app local sessions, supported OpenClaw records, and supported Cursor agent transcripts.

Can it recover cloud-only chats?

No. DataMoat preserves supported local records that are already present on your device and accessible to you.

Why not just export plaintext?

Plaintext exports are easy to leak, alter, or lose. DataMoat keeps encrypted records, skills folder snapshots, supported attachment blobs, and a hash-chained local audit log that can detect changed or broken entries in the current log.

How is my private data asset used?

DataMoat protects the asset first. You decide how protected work traces are later searched, exported, reused, handed off, or prepared for private memory.

What should I install first?

Use the macOS DMG on supported Macs, Windows ZIP packages on Windows, and source install on Linux or development machines. Windows users should extract the full ZIP folder and run Install DataMoat.cmd once.