DueFlow
中文 View on GitHub

Open source · local first · v0.1.1

Deadlines arrive scattered.
Your plan shouldn't.

DueFlow turns notices, screenshots, files, pasted text, and webhook payloads into structured tasks, reverse plans, risk checks, reminders, and calendar exports—on your machine.

Download for Apple Silicon Star DueFlow on GitHub Run the 60-second demo
  • Self-contained preview
  • No Python or Conda at runtime
  • MIT licensed
DueFlow dashboard showing Inbox, tasks, plans, and risk workflows
73project checks
3CI technology gates
0keys needed for demo
MITopen-source license

From noise to next action

One workflow for every deadline source.

Capture what arrived, understand what it means, and leave with a schedule you can edit and export.

  1. 01

    Collect

    Bring in text, Markdown, PDFs, OCR-ready images, file drops, local Inbox files, or webhooks.

  2. 02

    Extract

    Structure deadlines, deliverables, submission methods, missing details, and source references.

  3. 03

    Plan

    Generate editable tasks, reverse plans, risk checks, reminders, and priority-aware next steps.

  4. 04

    Export

    Move the result into Markdown reports, todo lists, summaries, or standard calendar files.

Local-first by default

Your deadlines belong to you.

DueFlow keeps its database, Inbox, exports, backups, and diagnostics on your machine. The deterministic mock provider lets you evaluate the complete workflow before configuring any external model provider.

Read the privacy boundary

Local storage

SQLite state, source files, backups, and generated exports stay under your configured paths.

Inspectable outputs

Plans and exports remain editable; source references make extracted results easier to review.

Recovery built in

Backup, restore, diagnostics, schema checks, and reproducible gates support safe local iteration.

Open implementation

Python, FastAPI, React, Tauri, and Rust components are available under the MIT license.

Try the complete workflow

From clone to generated plan in 60 seconds.

The default mock provider is deterministic, so the first run needs no account, token, or API key.

Expected output processed=3 tasks=3 plans=17 risks=5
Terminal 
$ git clone https://github.com/Ustinian5/DueFlow.git
$ cd DueFlow
$ conda env create -f environment.yml
$ conda run -n dueflow python scripts/run_demo.py

Apple Silicon developer preview

DueFlow v0.1.1 for macOS

Self-contained .app.zip with a loopback-only backend. Runtime needs no Python or Conda.

Download preview Verify SHA256 Release notes

Build calmer weeks

Make the deadline visible.
Make the next action obvious.

Explore the source, run the local demo, and help shape DueFlow's next release.

Star on GitHub Contribute