Nagi Tasks: a small browser-based task manager
This example embeds HTML/CSS/JavaScript and uses a Nagi JSON API with SQLite. It supports adding, editing, completing, deleting, filtering, and aggregate counts. It uses no external Web libraries or CDNs.
On this page
Start
Run from the repository root:
cargo build --release --locked -p nagic
.\target\release\nagic.exe run --project test-nagi-code/web-demo
Open http://127.0.0.1:8091 in your browser. NAGI_PORT changes the port; NAGI_DB changes the database path. Project mode defaults to test-nagi-code/web-demo/nagi-tasks.sqlite and retains data across restarts. The older run tasks.nagi form and direct executable execution use the starting working directory. To reuse an existing database, give its absolute path in NAGI_DB.
For a distributable Windows x64 executable:
.\scripts\build_windows_demo.ps1 -Source test-nagi-code/web-demo/tasks.nagi
.\build\distribution\nagi-tasks.exe
Use -Offline if dependencies are cached. HTML, SQLite, and the VC++ runtime are built into the exe; distribute that one executable. Saved user data lives in a separate SQLite file. The receiving machine does not need Rust, Nagi, or Python. Stop with Ctrl+C in the terminal.
Read the code
| File | Purpose |
|---|---|
| nagi.toml | Entry configuration shared by CLI and VS Code |
| tasks.nagi | HTTP handlers and startup |
| models.nagi | JSON and database types |
| validation.nagi | Input validation |
| index.html | Browser UI and API calls |
The UI is GET /. APIs are GET/POST /api/tasks, GET/PUT/DELETE /api/tasks/{id}, and GET /api/stats. Create/update bodies look like {"title":"A task","done":false}. Titles must be 1–240 UTF-8 bytes; the list shows the latest 100 tasks, while counts cover all tasks.
The example is a local loopback application. It does not include authentication or production hosting configuration.
Check an already running server
python test-nagi-code/web-demo/smoke_api.py --base-url http://127.0.0.1:8091
Python only sends HTTP requests and checks responses. It deletes the test data it created through the API and retains existing tasks. It neither starts/stops the server nor accesses database files.