Kodwai positions itself as the "LeetCode of the AI era," designed in response to changes in how programming ability is assessed after the spread of AI agents.
Purpose: According to its official description, Kodwai targets developer ability assessment scenarios. It argues that problems on traditional algorithm practice platforms (LeetCode) can be solved by agents in seconds, so grinding problems reflects memorization more than real engineering ability. Kodwai aims to measure developers' performance on real problems in a way that is closer to agent collaboration.
Main capabilities: Its official text describes its scoring target as the process in which developers collaborate with agents to complete real problems, including: pointing agents at problems, promptly detecting when agents are "confident but wrong," and checking the results the agents actually deliver. The platform says this ability is what determines a developer's level, and it is scored by Kodwai.
How to use: Officially, it only states that the assessment uses "real problems" and "your own agent," but it does not provide a specific operating process, registration method, problem format, or scoring criteria, and no detailed information is currently available.
Use cases: Suitable for developers who want to assess or demonstrate their ability to solve real programming problems with the assistance of AI agents, as well as teams or organizations concerned with measuring developer skills in the agent era.
Limitations: Public information is currently very limited. Apart from the above philosophy and positioning, no specific information is currently provided about pricing, functional modules, supported platforms, agent compatibility scope, assessment duration, or how results are presented.
