Build personal mini-tools in Qoder

Disclaimer: The content shown on this page is generated by an AI model and is for reference only. It does not constitute any guarantee of results. Actual output and Credits consumption are subject to "Usage Details" in the console.

This walkthrough uses a single-file HTML expense tracker to run a two-round Agent coding loop of "generate code → self-review and fix": first let the model generate code, then have it self-check boundary issues as a code reviewer and fix them, ultimately producing a ready-to-run expense tracker. Tool: Qoder · qwen3.7-max two-round coding loop.

Actual output

Run through with a single-file HTML expense tracker. The output interface (with sample data):

Click to try: Mini expense tracker

Step 1: One-sentence requirement

In Qoder, send the requirement below to qwen3.7-max:

Help me build a single-file HTML expense tracker:
Support recording date/category/amount/note, store data in localStorage,
display the total spending this month and a category pie chart at the top
(no external libraries, hand-draw with Canvas),
mobile browser friendly, fresh and soothing color scheme. Output a complete
runnable single-file HTML.

Step 2: Self-review and fix (key step in the Agent loop)

Please self-check the code above as a strict code reviewer: find functional defects
and boundary issues (such as invalid amount input, localStorage parsing failures,
pie chart display when there is no data), fix them, and output the final complete
single-file HTML. Only output the final version.

Notes

Consumption: This playbook involves a 2-round coding loop (generate code → self-review and fix). Round 2 requires reading back the entire code, so it is significantly more expensive than round 1.
Output: The final output is a 23.8K-character, directly runnable single-file expense tracker.

​Actual output

​Step 1: One-sentence requirement

​Step 2: Self-review and fix (key step in the Agent loop)

​Notes

Actual output

Step 1: One-sentence requirement

Step 2: Self-review and fix (key step in the Agent loop)

Notes