Ready-to-use disclosure badges and callouts for GitHub READMEs, based on the VisiData AI Contribution Levels.
Note
This project uses Level 0 AI assistance — no AI was used in the making of this code. Ethically-sourced, artisanal software written entirely by humans.
📋 Copy this markdown
## AI Contribution Disclosure
> [!NOTE]
> This project uses [Level 0 AI assistance](https://www.visidata.org/blog/2026/ai/) — no AI was used in the making of this code. Ethically-sourced, artisanal software written entirely by humans.Note
This project uses Level 1 AI assistance — AI was consulted for ideas, suggestions, or code review, but every character of code was written by a human. Think of it as using a chatbot instead of Stack Overflow.
📋 Copy this markdown
## AI Contribution Disclosure
> [!NOTE]
> This project uses [Level 1 AI assistance](https://www.visidata.org/blog/2026/ai/) — AI was consulted for ideas, suggestions, or code review, but every character of code was written by a human. Think of it as using a chatbot instead of Stack Overflow.Note
This project uses Level 2 AI assistance — a human wrote nearly all the code, with occasional copy-paste of small boilerplate snippets, regexes, or other minor incantations. The code was generated at the speed of human thought.
📋 Copy this markdown
## AI Contribution Disclosure
> [!NOTE]
> This project uses [Level 2 AI assistance](https://www.visidata.org/blog/2026/ai/) — a human wrote nearly all the code, with occasional copy-paste of small boilerplate snippets, regexes, or other minor incantations. The code was generated at the speed of human thought.Note
This project uses Level 3 AI assistance — a human wrote the code but incorporated some non-trivial AI-generated snippets or algorithms. The human treats this code as their own and takes full responsibility for it.
📋 Copy this markdown
## AI Contribution Disclosure
> [!NOTE]
> This project uses [Level 3 AI assistance](https://www.visidata.org/blog/2026/ai/) — a human wrote the code but incorporated some non-trivial AI-generated snippets or algorithms. The human treats this code as their own and takes full responsibility for it.Note
This project uses Level 4 AI assistance — a human wrote the majority of the code, but AI tools (such as Claude Code) were allowed to directly create or edit files, particularly for documentation, tests, and auxiliary artifacts.
AI Model: Claude Sonnet 4
📋 Copy this markdown
## AI Contribution Disclosure
> [!NOTE]
> This project uses [Level 4 AI assistance](https://www.visidata.org/blog/2026/ai/) — a human wrote the majority of the code, but AI tools (such as Claude Code) were allowed to directly create or edit files, particularly for documentation, tests, and auxiliary artifacts.
>
> **AI Model:** Claude Sonnet 4Important
This project uses Level 5 AI assistance — AI generated the majority of the code, but the human was involved at every step, reviewing each line with full attention. The human understands every algorithm and has logically validated how it all works.
AI Model: Claude Opus 4
📋 Copy this markdown
## AI Contribution Disclosure
> [!IMPORTANT]
> This project uses [Level 5 AI assistance](https://www.visidata.org/blog/2026/ai/) — AI generated the majority of the code, but the human was involved at every step, reviewing each line with full attention. The human understands every algorithm and has logically validated how it all works.
>
> **AI Model:** Claude Opus 4Important
This project is an experiment in Level 6 AI-assisted coding — AI does most of the coding while the human acts as team lead, driving development direction and testing empirically. The author is a seasoned software engineer who steers architecture and reviews results, but defers much of the implementation to AI tooling. This is not production-ready software; it's a hobby project and a learning exercise in pushing the boundaries of AI-assisted development.
AI Model: Claude Opus 4
📋 Copy this markdown
## AI Contribution Disclosure
> [!IMPORTANT]
> This project is an experiment in [Level 6 AI-assisted coding](https://www.visidata.org/blog/2026/ai/) — AI does most of the coding while the human acts as team lead, driving development direction and testing empirically. The author is a seasoned software engineer who steers architecture and reviews results, but defers much of the implementation to AI tooling. This is not production-ready software; it's a hobby project and a learning exercise in pushing the boundaries of AI-assisted development.
>
> **AI Model:** Claude Opus 4Warning
This project uses Level 7 AI assistance — the human specified functionality at a high level and let AI generate the implementation. The human focused on tests, specs, and process rather than the code itself, acting more like a manager than an individual contributor. Code has been empirically validated but not line-by-line reviewed.
AI Model: Claude Code with Opus 4
📋 Copy this markdown
## AI Contribution Disclosure
> [!WARNING]
> This project uses [Level 7 AI assistance](https://www.visidata.org/blog/2026/ai/) — the human specified functionality at a high level and let AI generate the implementation. The human focused on tests, specs, and process rather than the code itself, acting more like a manager than an individual contributor. Code has been empirically validated but not line-by-line reviewed.
>
> **AI Model:** Claude Code with Opus 4Warning
This project uses Level 8 AI assistance — AI worked mostly autonomously on features and bugs while the human acted as an executive, spending minimal time on basic validation and review. Beyond smoke testing, the code is trusted to be correct enough.
AI Model: Claude Code (autonomous mode)
📋 Copy this markdown
## AI Contribution Disclosure
> [!WARNING]
> This project uses [Level 8 AI assistance](https://www.visidata.org/blog/2026/ai/) — AI worked mostly autonomously on features and bugs while the human acted as an executive, spending minimal time on basic validation and review. Beyond smoke testing, the code is trusted to be correct enough.
>
> **AI Model:** Claude Code (autonomous mode)Caution
This project uses Level 9 AI assistance — a human instructed AI to act autonomously on this work and gave it no further attention. Use at your own risk.
AI Model: Autonomous AI Agent
📋 Copy this markdown
## AI Contribution Disclosure
> [!CAUTION]
> This project uses [Level 9 AI assistance](https://www.visidata.org/blog/2026/ai/) — a human instructed AI to act autonomously on this work and gave it no further attention. Use at your own risk.
>
> **AI Model:** Autonomous AI AgentCaution
This project uses Level 10 AI assistance — no human instigated or oversaw this work. This was generated by autonomous bots acting on their own initiative.
📋 Copy this markdown
## AI Contribution Disclosure
> [!CAUTION]
> This project uses [Level 10 AI assistance](https://www.visidata.org/blog/2026/ai/) — no human instigated or oversaw this work. This was generated by autonomous bots acting on their own initiative.| Level | Description | Badge Color | Human Role |
|---|---|---|---|
| 0 | No AI used | blue |
Solo developer |
| 1 | Ideas only | blue |
Developer + rubber duck |
| 2 | Minor assists | cyan |
Developer + autocomplete |
| 3 | Non-trivial assists | green |
Developer + copilot |
| 4 | Significant help | yellow |
Lead developer |
| 5 | Bots code, human understands all | orange |
IC with full review |
| 6 | Bots code, human understands most | red |
Team lead |
| 7 | Human specs, bots implement | pink |
Manager |
| 8 | Bots plan, human approves | purple |
Executive |
| 9 | Fire and forget | grey |
Absent |
| 10 | Rogue bots | black |
None |
Based on VisiData's AI Contribution Levels by Saul Pwanson.