prompt
FreeUltra-compressed agent communication with full technical accuracy
FreeFree tier
About prompt
Caveman Mode is an ultra-compressed communication prompt for AI assistants, designed to reduce output token usage by approximately 75% while preserving full technical accuracy. Developed by Julius Brussee and hosted in the ai-boost/awesome-prompts repository, this prompt instructs AI to respond like a 'smart caveman'—dropping filler words, articles, and pleasantries while keeping all technical substance. It includes four intensity levels (lite, full, ultra, wenyan) and an auto-clarity feature for safety-critical messages. Ideal for developers and power users seeking efficient, no-nonsense interactions in debugging, code explanation, and technical discussions.
Key Features
~75% output token reduction while retaining technical accuracy
Four intensity levels: lite, full, ultra, wenyan (classical Chinese terseness)
Auto-clarity mode for security warnings, destructive ops, and multi-step sequences
Rules to drop articles, filler words, pleasantries, and hedging
Preserves code blocks, file paths, commands, and error messages unchanged
Pattern-based responses: [thing] [action] [reason] → [next step]
Level persists until changed or session ends
Pros & Cons
Pros
- Significant token reduction saves on API costs and latency
- Maintains technical precision and accuracy
- Customizable intensity to suit different contexts
- Auto-clarity ensures safety for irreversible operations
- Open-source and free to use
- Clear, unambiguous instructions for AI adherence
Cons
- May come across as rude or unnatural in user-facing communication
- Requires context switching for normal prose in code/commits
- Not suitable for casual or conversational interactions
- Effectiveness depends on AI model compliance with prompt rules
- Limited to text-based outputs; no multimedia support
Best For
Technical support chat and debugging sessionsCode explanation and system design discussionsDocumentation generation and reviewEfficient communication in CI/CD pipelines or bot responsesReducing API costs by compressing AI output tokensRapid troubleshooting in development environments
FAQ
What is Caveman Mode?
Caveman Mode is a prompt that instructs AI to communicate in an ultra-compressed style, reducing output tokens by about 75% while keeping all technical substance intact.
How do I use it?
Copy the prompt from the GitHub file and use it as a system instruction with your AI assistant. You can specify an intensity level (lite, full, ultra, wenyan) at the start.
What are the different intensity levels?
Lite removes filler and hedging but keeps articles and full sentences. Full drops articles and uses fragments. Ultra abbreviates and strips conjunctions. Wenyan uses classical Chinese terseness.
Does it affect code or commands?
No. Code blocks, file paths, commands, and error messages remain unchanged. The compression applies only to explanatory or chat text.
What is auto-clarity?
Auto-clarity temporarily disables caveman mode for security warnings, irreversible action confirmations, multi-step sequences, or when the user asks for clarification. It resumes caveman after the clear part is communicated.