How We Score AI Coding Tools
Updated April 28, 2026 · Data as of 2026-04-26
Every tool and model in the WhichAI catalog is scored independently by Erik Lewis based on direct use, vendor documentation, and community feedback. No tool pays to be included, and no tool's position changes based on affiliate relationships. The scoring engine produces the same result regardless of whether a tool has an affiliate program.
How we score
Each tool is rated 0 to 5 on 11 dimensions. The scale is anchored: 0 means the tool has no meaningful capability in that dimension, and 5 means it is among the best available. Scores reflect the tool as it stands at the time of the last update — not roadmap promises.
| Dimension | What it measures (0 = worst, 5 = best) |
|---|---|
| beginner friendly | Ease of use for someone with no prior AI tool experience (0=experts only, 5=zero barrier) |
| free tier quality | How useful is the free offering? (0=no free tier, 5=fully functional free forever) |
| agentic power | Autonomous multi-step coding capability (0=autocomplete only, 5=full repo-aware agent) |
| vibe coding fit | Natural language → working app capability (0=not designed for this, 5=purpose-built) |
| ide integration | Fits into existing IDE workflow (0=no integration, 5=seamless native) |
| multi file editing | Cross-file and repo-wide editing (0=single line, 5=full repo orchestration) |
| community resources | Tutorials, docs, community size (0=no community, 5=massive ecosystem) |
| data analysis fit | Suitability for data science and notebook work (0=not suited, 5=purpose-built) |
| privacy safety | Data handling safety (0=significant known concerns, 5=enterprise-grade) |
| google ecosystem | Integration with Google Workspace and Cloud (0=none, 5=native deep integration) |
| microsoft ecosystem | Integration with Microsoft 365, GitHub, Azure (0=none, 5=native deep integration) |
How quiz weights work
Your quiz answers determine which dimensions matter most for your situation. A beginner who wants free tools gets a result weighted heavily toward beginner-friendliness and free tier quality. An experienced developer building complex apps gets a result weighted toward agentic power and multi-file editing. The engine normalizes the weights from your answers and ranks all eligible tools. The highest-scoring tool for your specific weight set becomes your recommendation.
Update cadence
The full catalog is reviewed weekly during active market movement — in 2026, tools are changing pricing and releasing major updates monthly. Individual entries update immediately when a significant vendor change is detected: a new pricing tier, a renamed plan, a deprecation, or a shutdown. Subscribers receive an email when the best fit for their specific quiz answers changes.
About the author
Erik Lewis is an independent researcher and builder based in West Plains, Missouri. He got serious about AI coding tools when the agentic wave took hold and has tracked every meaningful shift in the market since. He is not affiliated with any AI company or tool vendor. He runs TheAIConvert, where he covers the broader AI adoption landscape for practitioners.
Funding and disclosure
This site earns revenue through Google AdSense display advertising. Ads appear in passive zones only — quiz question screens, answer buttons, and your recommendation card are always clean. We also participate in affiliate programs for some tools listed, where we earn a small commission if you sign up through a link on this site. Affiliate status does not affect scores or rankings. The scoring engine produces the same result whether or not a tool has an affiliate program.
Contact
Questions or corrections: hello@whichai.tech or visit the Contact page.