Add ActivityWatch-compatible API and database layer

[mk1316]

Summary

This PR introduces ActivityWatch (AW) compatibility to the application by adding a new database layer and API commands that follow the ActivityWatch data model. This enables the app to store and query activity data in a format compatible with ActivityWatch tools and integrations.

Key Changes

Database & Models

• New migration (0003_activitywatch_compat.sql): Creates ActivityWatch-compatible schema with:

  • buckets table: Containers for events from specific watchers on specific hosts
  • events table: Activity data with flexible JSON payloads
  • key_value table: Settings and metadata storage
  • Views for easy querying (window events, usage summaries)
  • Migration of existing session data to new event format

• New models (aw_models.rs): ActivityWatch-compatible data structures:

  • Bucket: Container for events from a watcher
  • Event: Activity event with timestamp, duration, and JSON data
  • Heartbeat: Efficient event submission format
  • ServerInfo: API info response
  • Helper types for window, AFK, web tab, and editor events

• New database layer (aw_database.rs): Async database operations including:

  • Bucket CRUD operations
  • Event management with filtering and pagination
  • Heartbeat processing with intelligent event merging (extends events within pulsetime if data matches)
  • Usage summary queries
  • Key-value store for settings

API Commands

• New commands (aw_commands.rs): 20+ Tauri commands exposing the ActivityWatch API:
  • Info: aw_get_info
  • Buckets: aw_get_buckets, aw_get_bucket, aw_create_bucket, aw_delete_bucket
  • Events: aw_get_events, aw_get_event, aw_insert_events, aw_delete_event, aw_get_event_count
  • Heartbeat: aw_heartbeat (efficient event submission)
  • Queries: aw_get_usage_summary, aw_get_current_event
  • Settings: aw_get_setting, aw_set_setting
  • Export: aw_export_bucket, aw_export_all

Integration

• Tracking integration: Modified start_tracking() to:

  • Initialize ActivityWatch bucket for window tracking
  • Send heartbeats on app switches and during idle periods
  • Maintain backward compatibility with legacy session tracking

• Dependencies: Added gethostname crate for hostname detection

Notable Implementation Details

• Heartbeat merging: The heartbeat endpoint intelligently merges consecutive events with identical data that occur within the pulsetime window, reducing storage overhead
• Backward compatibility: Existing session data is automatically migrated to the new event format
• Flexible data model: Events use JSON for the data field, allowing arbitrary activity types
• Async/await: All database operations are async using sqlx with tokio runtime
• Persistent device ID: Reuses existing device ID system for consistent identification

Testing

The database layer includes unit tests for datetime parsing to ensure compatibility with both RFC3339 and SQLite datetime formats.

https://claude.ai/code/session_01QYNePFqXFj4AaRtTDJVe6J

---

Note

High Risk
Large cross-cutting change touching persistence/migrations, background tracking, a new local HTTP server, and cloud sync plumbing; regressions could affect data integrity, performance, or expose unintended local API behavior.

Overview
Adds an ActivityWatch-compatible data layer to the desktop app: a new SQLite migration introduces buckets, events, and key_value tables (plus indexes/views) and migrates legacy sessions into AW-style events.

Introduces new Rust modules for AW models, database operations (including heartbeat merging), Tauri command handlers (CRUD, export, settings, summaries), and an Axum REST server exposing AW-like endpoints on 127.0.0.1:5600; the app startup now enables SQLite foreign keys for all pooled connections and manages an additional AwDb state.

Updates usage tracking to create an AW window bucket and continuously emit AW heartbeats (on app switches and periodic keep-alives) while keeping legacy session tracking.

Adds a Neon-based cloud sync implementation (neon.rs) and updates frontend hooks/types to support AW APIs, AW query/categorization helpers, and Neon bucket/event sync; Supabase sync is marked deprecated. Webclient also includes small auth hardening (null-supabase guard) and UI/theming updates (fonts + global CSS refresh).

^{Written by Cursor Bugbot for commit 53e3ea9. This will update automatically on new commits. Configure here.}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
loopd	Error		Feb 6, 2026 9:16pm

[cursor]

Unused delete_setting method is dead code

Low Severity

The delete_setting method in AwDatabase is defined but never called. There's no Tauri command or REST API endpoint that exposes this functionality. The key-value store API only has get_setting and set_setting exposed to consumers.

 

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with Cloud Agents, enable Autofix in the Cursor dashboard.}