Skip to main content

AI Coding · Add features to an existing system

Describe the requirement in one sentence, and the AI reads the existing code to incrementally add features without touching the old logic.

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 sales CRM: it lacks a "customer follow-up reminder" feature, so overdue customers go unattended. In Qoder, you hand the existing code to qwen3.7-max and, with one sentence, ask it to incrementally add the reminder without touching old features such as search and filtering. Tool: Qoder · qwen3.7-max reads the existing code → incremental development.

Actual output

Same existing codebase, only one feature instruction added. Before and after adding the feature (both are clickable to try):

Basic CRM before adding the feature

Basic CRM

Click to try

After adding the follow-up reminder

Follow-up reminder added

Click to try

Steps: Add a feature with one sentence

In Qoder, send the existing code along with the instruction below to qwen3.7-max:
This is my existing CRM system code (see below). Please add a "customer follow-up reminder"
feature without modifying the existing features (search, status filtering):
1. Customers not followed up for more than 7 days should be highlighted in red and pinned
   to the top (closed and lost customers are excluded from reminders);
2. Display a statistics bar at the top of the page: "Follow-up reminders: N customers
   not followed up for more than 7 days";
3. Add a "Days since last follow-up" column that shows the number of days, with reminder
   customers highlighted in red bold;
Keep the existing code style and output the complete modified file.
[Existing code] {paste file content}

Notes

  • Consumption: This playbook involves a single round of incremental development (within the Qoder tool, requires reading the full existing code). Consumption in brownfield development is positively correlated with codebase size — the existing code must enter the context, and larger code means higher single-round input.
  • Consistency: The feature passed on the first try: overdue items correctly pinned and highlighted in red, closed and lost customers correctly excluded, and exactly 7 days did not trigger a false alarm.
Model Playground
Fine-tuning
Model Compression
Statistics and Monitoring
Support