kodwai
Kodwai is a free platform for AI-agent coding challenges that you solve locally with your own tools (e.g., Claude Code, Cursor, Codex), and it scores how well you direct, verify, and ship with an agent—not what you memorized.
https://kodwai.com/?ref=producthunt
Product Information
Updated:Jun 29, 2026
What is kodwai
Kodwai is a developer-focused challenge platform designed for the “vibe coding” era, where engineers build software by collaborating with AI agents. Instead of testing puzzle-solving ability, Kodwai evaluates real, ticket-sized coding work completed on your own machine using your preferred agent and editor. After you submit, Kodwai generates a score and public profile that reflects not only whether the solution works, but also the quality of your process—capturing prompts, commits, test runs, and recovery when an agent is wrong.
Key Features of kodwai
Kodwai is a free, local-first coding challenge platform designed for the “AI-agent” era: you solve real, ticket-sized engineering problems on your own machine using your preferred AI agent (e.g., Claude Code, Cursor, Codex), and Kodwai scores the entire session—not just whether tests pass. It evaluates how well you direct and verify the agent via three axes (Direction, Outcome, Lift), using evidence from prompts/transcripts, commits, test runs, and timing, then publishes results to a leaderboard and a shareable developer profile with ranks and badges.
AI-agent session scoring (Direction / Outcome / Lift): Scores how you steer, verify, decompose, and recover with an AI agent (Direction), what actually shipped and whether it holds (Outcome), and the extra edge-case rigor beyond a one-shot prompt (Lift), with per-signal evidence.
Local-first workflow via CLI: Challenges run on your own machine (no browser sandbox). The CLI downloads PROBLEM.md, starter files, and tests, initializes a git repo, starts the timer, and later submits the full run for scoring.
Bring-your-own agent support: Works with popular agents like Claude Code and Cursor as first-class options, and supports other terminal-based agents (e.g., Codex CLI, Aider, Cline), letting developers use their real setup.
Real-world, ticket-sized challenges: Problems are scoped like practical engineering tasks rather than riddles, spanning multiple categories and difficulties, aimed at reflecting how developers actually ship software.
Evidence-backed evaluation (not just green tests): Submission packages code, git history, test runs, agent transcript, and time; scoring cites specific evidence (e.g., transcript turns or verification steps) to explain why you earned points.
Leaderboards, profiles, and auto-earned badges: Each scored run affects a difficulty-weighted global leaderboard and builds a public profile showing score breakdown, rank, badges (milestones/streaks/skill/agent), and the agents you used.
Use Cases of kodwai
Developer skill benchmarking in the AI era: Individuals can measure and improve their real-world ability to collaborate with AI agents (prompting, verification, recovery) rather than practicing memorization-heavy puzzle formats.
Hiring portfolios and candidate signaling: Developers can share a public profile (score, rank, badges, agent usage) with recruiters/hiring managers as an alternative signal to take-home projects or LeetCode-style screens.
Team upskilling and AI workflow training: Engineering teams can use challenges to practice safe, verifiable agent-driven development habits—writing specs, adding tests, probing edge cases, and preventing “tests-green-but-wrong” outcomes.
Tooling comparison across AI agents: Developers can run similar challenges with different agents (Claude Code vs Cursor vs Codex, etc.) to understand which workflows produce better Direction/Lift and more reliable shipping outcomes.
Competitive learning and motivation: Leaderboards, streaks, and badges create a structured way to stay consistent, track progress over time, and compete with peers while building practical agent-collaboration skills.
Pros
Measures real agent-driving skill (steering, verification, recovery), not just whether tests pass.
Local-first and bring-your-own-agent mirrors real development environments and workflows.
Transparent, evidence-cited scoring and a shareable public profile/leaderboard for signaling progress.
Cons
Requires installing/using a CLI and running locally, which may add friction for users who prefer browser-only platforms.
Scoring depends on capturing transcripts/commits/test runs; users may have privacy concerns about what gets submitted.
Best value assumes you regularly use AI agents; developers who don’t may find the premise less relevant.
How to Use kodwai
1) Choose a challenge on Kodwai: Go to https://kodwai.com/ and browse the available challenges. Pick one that matches the category and difficulty you want to practice (they’re scoped like real, ticket-sized problems).
2) Prepare your local setup (BYO agent): Decide which AI coding agent you’ll use on your own machine (e.g., Claude Code, Cursor, Codex CLI, or another terminal-driven agent). Kodwai is designed for “bring your own agent” workflows.
3) Start the challenge with the Kodwai CLI: In your terminal, run the Kodwai CLI to download the challenge materials (PROBLEM.md, starter files, and tests), initialize a git repository, and start the timer: `$ npx @kodwai/cli challenge <slug>` where `<slug>` is the challenge identifier from the site.
4) Read the problem and constraints locally: Open the downloaded `PROBLEM.md` and any starter code in your editor. Make sure you understand the requirements, edge cases, and what the tests are asserting.
5) Solve locally using your agent + your editor: Work on the solution on your own machine (no browser sandbox). Use your AI agent to help implement, refactor, and reason—but steer it actively: clarify the spec, break the work into steps, and verify assumptions.
6) Run and re-run the provided tests as you iterate: Execute the included test suite locally throughout development to confirm correctness. If tests fail, use the feedback to guide fixes and improvements.
7) Commit meaningful progress to git: Since Kodwai scores the whole session (including git history), make commits as you reach logical milestones (e.g., initial implementation, bug fix, edge-case handling, concurrency fix).
8) Add verification and edge-case coverage: Don’t rely on a one-shot prompt. Strengthen the solution by probing tricky cases (e.g., concurrency, boundary conditions, performance constraints) and, when appropriate, add or adjust tests to prove key claims.
9) Submit your run from the CLI: When you’re satisfied and tests pass, submit with: `$ npx @kodwai/cli submit`. This packages your code, git history, test runs, agent transcript, and time taken for scoring.
10) Review your score and evidence: After submission, Kodwai returns a score across three axes—Direction, Outcome, and Lift—with per-signal evidence drawn from your transcript, commits, and test runs so you can see exactly why you scored what you did.
11) Check the leaderboard and your public profile: View your placement on the public leaderboard and your developer profile (shown as `kodwai.com/developers/you` in the site copy). Your profile reflects your scores, rank, badges, and which agents you used.
12) Improve by running more challenges: Repeat with additional challenges to climb the leaderboard and improve the parts Kodwai emphasizes most—especially Direction (steering, verification, decomposition, and recovery when the agent is wrong).
kodwai FAQs
Kodwai is a platform for AI-agent coding challenges where developers solve real, ticket-sized problems on their own machine using their preferred AI agent (e.g., Claude Code, Cursor, Codex). It scores how well you direct the agent during the session and ranks you on a public leaderboard.
kodwai Video
Popular Articles
Atoms: A Multi-Agent AI Platform That Transforms Ideas into Launch-Ready Products
May 22, 2026
Nano Banana SBTI: What It Is, How It Works, and How to Use It in 2026
Apr 15, 2026
Atoms Review — The AI Product Builder Redefining Digital Creation in 2026
Apr 10, 2026
Kilo Claw: How to Deploy and Use a True "Do‑It‑For‑You" AI Agent(2026 Update)
Apr 3, 2026
Latest AI Tools Similar to kodwai
Gait
Gait is a collaboration tool that integrates AI-assisted code generation with version control, enabling teams to track, understand, and share AI-generated code context efficiently.
invoices.dev
invoices.dev is an automated invoicing platform that generates invoices directly from developers' Git commits, with integration capabilities for GitHub, Slack, Linear, and Google services.
EasyRFP
EasyRFP is an AI-powered edge computing toolkit that streamlines RFP (Request for Proposal) responses and enables real-time field phenotyping through deep learning technology.
Cart.ai
Cart.ai is an AI-powered service platform that provides comprehensive business automation solutions including coding, customer relations management, video editing, e-commerce setup, and custom AI development with 24/7 support.
Popular AI Tools Like kodwai
GitHub Copilot Chat
GitHub Copilot Chat is an AI-powered coding assistant that provides natural language interactions, real-time code suggestions, and contextual support directly within supported IDEs and GitHub.com.
CopilotForXcode
CopilotForXcode is an Xcode Source Editor Extension that integrates GitHub Copilot, Codeium, and ChatGPT to provide AI-powered code suggestions, chat assistance, and prompt-to-code functionality within Xcode.
BrowserAI
BrowserAI is an open-source library that enables running local Large Language Models (LLMs) directly in web browsers with WebGPU acceleration, offering privacy-focused AI capabilities without requiring server infrastructure.
OpenAI Codex CLI
OpenAI Codex CLI is a lightweight, open-source coding agent that runs in your terminal, enabling developers to translate natural language into code execution while providing ChatGPT-level reasoning with the ability to run code, manipulate files, and iterate under version control.