Data Storage¶

All Lumbergh data lives in a single directory. No database server required.

Data Directory¶

Default location: ~/.config/lumbergh/

Override with the LUMBERGH_DATA_DIR environment variable.

File Layout¶

~/.config/lumbergh/
├── sessions.json              # Session registry (names, workdirs, agent, the_one flag, scratch flag)
├── projects/
│   └── <repo-hash>.json       # Per-project data (todos, scratchpad, prompts, AI prompts)
├── session_data/
│   └── <session-name>.json    # Per-session runtime data (idle state, status, AI summaries)
├── settings.json              # AI provider config, repo search path, git graph commits, password,
│                              # cloudToken, telemetryConsent, scratchMaxAgeDays, backupPassphrase
├── global.json                # Global prompt templates and global AI prompts
├── shared/                    # Shared files accessible from all sessions
└── scratch/
    └── <name>/                # Throwaway worktrees for scratch sessions (auto-cleaned)

File	Contents
`sessions.json`	Session registry — names, working directories, descriptions, agent provider, `the_one` star flag (exposed as `theOne` over the API), session type (e.g. `scratch`), display names, paused/cloud flags, `lastUsedAt` timestamp
`projects/<hash>.json`	Per-project todos, scratchpad notes, project-specific prompts, and AI prompt overrides. Keyed by repo path hash, so multiple sessions on the same repo share data.
`session_data/<name>.json`	Per-session runtime data — idle state, status, AI-generated session summaries. One file per session.
`settings.json`	Global settings — AI provider configuration, repo search path, git graph commit count, default agent, tab visibility, cloud token, telemetry consent, scratch max-age, password, backup passphrase
`global.json`	Global prompt templates and global AI prompt templates
`shared/`	Shared files (including images) accessible from every session
`scratch/<name>/`	Working directories for scratch sessions. Auto-deleted after `scratchMaxAgeDays` of inactivity.

Format¶

All files are human-readable JSON stored via TinyDB. You can safely inspect and hand-edit them with any text editor.

Warning

Edit data files only while Lumbergh is stopped, or you risk write conflicts.

Backup & Restore¶

Manual Backup¶

Back up everything by copying the data directory:

cp -r ~/.config/lumbergh/ ~/lumbergh-backup/

To restore, copy the backup back to the same path (or point LUMBERGH_DATA_DIR at it).

Cloud Backup¶

With a Lumbergh Cloud connection, you can enable automatic cloud backup in Settings --> Cloud. Backups run every 5 minutes when changes are detected and include sessions, todos, prompts, scratchpads, shared markdown files, and settings (with secrets stripped by default). Backups support optional AES-256 encryption with a passphrase. Restore from the Cloud tab to pull your data to any machine.