kodwai
AI-agent coding challenges that score how well you direct them
Freemium - Free tier available for basic challenges; paid/pro plans and team features for advanced usage (see site for exact pricing).
About kodwai
kodwai offers real coding challenges you solve by directing your own AI agent. Challenges are test-driven and kodwai scores submissions across Direction (steering & verification), Outcome (correctness), and Lift (improvements/edge probing). Session-based prompts, one-shot or iterative runs, and automated test feedback help developers practice prompt engineering, agent orchestration, and automated verification. Visible examples show integration with Anthropic Claude models.
Key Features
- Run real, test-driven coding challenges using your own AI agent
- Automated scoring across Direction, Outcome, and Lift to evaluate prompting skill
- Session-based prompts with one-shot and iterative runs plus test feedback
- Supports LLMs such as Anthropic Claude and focuses on verification and agent steering
Who is this for?
Developers learning to build and direct AI agents, hiring teams assessing coding/agent skills, and engineers practicing LLM-driven workflows.
Comments (0)
No comments yet
Be the first to share your thoughts!
Sign in to join the conversation
Sign InStats
Categories
Pricing
FreemiumFree tier available for basic challenges; paid/pro plans and team features for advanced usage (see site for exact pricing).
Platforms
Tech Stack
Tags
Related Products
Claude Code
AI pair programmer that understands your entire codebase
GitHub Copilot
Your AI pair programmer from GitHub
Dodo Payments
Seamless payments and billing for AI, SaaS, and digital products.
DeepSeek Chat
Open-source AI that rivals GPT-4 at fraction of the cost
Scrapfly
Effortless web data extraction with AI-powered APIs.