vibe-budget launches CLI to forecast LLM coding costs

Cost predictability is becoming a core feature for AI coding workflows, and vibe-budget turns that into a practical preflight check.

• –Task-level input/output token breakdowns make budget trade-offs visible before prompts start
• –Side-by-side comparisons across 85+ models reduce the chance of accidentally picking expensive defaults
• –The scan command is useful for planning refactors and feature extensions on existing codebases
• –Accuracy depends on OpenRouter price freshness and how well model quality heuristics match real tasks