Self-hosted · Open source · MIT

Your tasks, finally
in one place.

Task Sync bridges Google Tasks and Microsoft To Do with field-level conflict resolution. Runs on your machine. Your data never leaves.

Google Tasks ⇄ Microsoft To Do

View on GitHub Get started

Runtime deps

Sync fields

Sync modes

Data leaks

Google Tasks on the phone.
Microsoft To Do at work.
Same tasks. Two silos. Zero sync.

⚠

Tasks fall through the cracks

You add a task on your phone and forget to add it at work. Or vice versa. Things get missed.

🔒

Third-party services want your data

Cloud sync tools route your tasks through their servers. Your data becomes their product.

💥

Whole-task sync causes data loss

Simple sync tools overwrite entire tasks on conflict. Change the title here, lose the due date there.

Features

Built for people who care about their data

Every feature exists because something annoyed us about existing tools.

⚔

Field-level conflict resolution

Change the title in Google Tasks. Change the due date in Microsoft To Do. Both changes survive. Unlike task-level sync, every field is compared individually — last-write-wins per field, not per task. Conflicts logged for transparency.

🔍

Cold-start matching

First sync won't duplicate everything. Tasks are matched across providers by title + notes before creating anything new.

🛡

Self-hosted

Runs on your machine. Your OAuth tokens, your state files, your data. No cloud middleman. No telemetry.

🗑

Delete propagation

Tombstones ensure deleted tasks stay deleted across providers. 30-day TTL prevents ghost resurrections.

🎮

Two interfaces, one engine

Use the Web UI for one-click OAuth and visual sync controls with auto-sync timers. Or use the CLI for cron jobs, scripts, and headless environments with JSON output. Same sync engine underneath.

⚡

Lean runtime

Only commander and zod at runtime. Tiny attack surface.

How It Works

The sync pipeline

Every sync run is an 11-step pipeline that ensures nothing gets lost.

Lock

Acquire file lock

Load

Read state & mappings

Prune

Expire old tombstones

Fetch

Pull tasks from APIs

Match

Cold-start deduplicate

Diff

Field-by-field compare

Resolve

Last-write-wins

Push

Fan out to providers

Save

Persist new state

Get Started

Running in under a minute

Choose your interface. Same sync engine underneath.

Web UI

# Clone & install

$ git clone https://github.com/salaamdev/task-sync.git

$ cd task-sync && npm install

$ npm run web:install

# Configure OAuth credentials

$ cp .env.local.example .env.local

# Edit .env.local with your keys

# Launch

$ npm run web:dev

▸ Ready on http://localhost:3000

Connect providers → Sync Now → Done.

CLI

# Build

$ git clone https://github.com/salaamdev/task-sync.git

$ cd task-sync && npm install && npm run build

# Get OAuth tokens

$ npm run oauth:google

$ npm run oauth:microsoft

# Sync

$ node dist/cli.js doctor

✓ Config valid ✓ Google ✓ Microsoft

$ node dist/cli.js sync

✓ Synced 23 tasks (2 created, 1 updated)

# Auto-sync every 5 min

$ node dist/cli.js sync --poll 5

⏳ Next sync in 5:00

Web UI Dashboard Task Sync Web UI showing connected Google Tasks and Microsoft To Do with sync report

Sync Modes

Your data, your rules

Three modes for different workflows. Set once in your config.

⇄

Bidirectional

Changes flow both ways. Edit anywhere, sync everywhere.

default

→

A → B Only

One-way push. Provider A is the source, B receives updates.

one-way

⊜

Mirror

Provider A is truth. B becomes an exact replica. Additions, edits, deletes — all mirrored.

strict

Your tasks, finally
in one place.

Tasks fall through the cracks

Third-party services want your data

Whole-task sync causes data loss

Built for people who care about their data

Field-level conflict resolution

Cold-start matching

Self-hosted

Delete propagation

Two interfaces, one engine

Lean runtime

The sync pipeline

Lock

Load

Prune

Fetch

Match

Diff

Resolve

Push

Save

Running in under a minute

Your data, your rules

Bidirectional

A → B Only

Mirror

Lean stack, zero bloat

One command. Both task lists. Always in sync.

Your tasks, finallyin one place.

Tasks fall through the cracks

Third-party services want your data

Whole-task sync causes data loss

Built for people who care about their data

Field-level conflict resolution

Cold-start matching

Self-hosted

Delete propagation

Two interfaces, one engine

Lean runtime

The sync pipeline

Lock

Load

Prune

Fetch

Match

Diff

Resolve

Push

Save

Running in under a minute

Your data, your rules

Bidirectional

A → B Only

Mirror

Lean stack, zero bloat

One command. Both task lists. Always in sync.

Your tasks, finally
in one place.