Llm Cost Optimizer
Reduce LLM API costs by 30-60% through prompt compression, caching, and model optimization
What You Can Do
You can identify inefficiencies in your LLM API spending and receive prioritized recommendations with validated ROI estimates. The skill performs static analysis on your current prompts and configurations, examines dynamic usage patterns to pinpoint high-cost workflows, and runs comparative modeling across optimization scenarios. You'll get implementation roadmaps that quantify financial impact for each change.
Features
Identifies redundant tokens and restructures prompts to maintain quality while reducing input size by 20-40%
Recommends where prompt caching delivers maximum ROI based on your repetition patterns and API pricing
Analyzes task complexity and suggests cheaper model alternatives (e.g., Claude 3.5 Haiku vs. Opus) with quality tradeoff assessment
Identifies opportunities to group requests and reduce per-call overhead by 15-35%
Analyzes 2+ weeks of API logs to surface cost drivers and high-impact optimization targets
Quantifies monthly and annual savings for each optimization, prioritized by implementation effort
Provides step-by-step guidance for deploying changes without breaking production systems
Establishes metrics to ensure optimizations maintain output quality and performance standards
Example Output
Example 1: Prompt Compression
Original prompt: 1,200 tokens → Optimized: 720 tokens
CompressionRatio: 40% reduction
Method: Removed redundant context, converted instructions to structured format
MonthlySavings: $180 (at 0.003/1K tokens)
Example 2: Model Tier Switch
Current: 10,000 API calls/month using Claude 3.5 Opus
Recommendation: Switch 60% of calls to Claude 3.5 Haiku (for classification/extraction tasks)
ProjectedSavings: $420/month (42% reduction in token costs)
QualityRisk: Low — Haiku scores 94% accuracy on your existing test set
Example 3: Caching Impact
High-repetition workflow identified: 500 calls/week with identical 300-token system prompt
Caching opportunity: Cache the system prompt + conversation history
Savings: 150,000 tokens/month × $0.0008 = $120/month
Implementation effort: 30 minutes (add caching headers to API calls)
What's Included
- SKILL.md: Complete optimization framework with decision trees for each optimization vector
- Cost Analysis Template: Spreadsheet to input your API usage data and calculate baselines
- Prompt Compression Checklist: Step-by-step guide to identify and remove redundant tokens
- Model Selection Matrix: Capability and cost comparison chart across Claude models (3.5 Haiku, Sonnet, Opus)
- Implementation Roadmap Template: Prioritized action plan with effort estimates and safety checkpoints
- ROI Validation Worksheet: Track pre/post-optimization costs to measure actual savings
Who It's For
- AI/Prompt Engineers — Responsible for prompt design and API integration optimization
- ML/AI Team Leads — Managing LLM infrastructure costs at scale
- Finance/Operations Managers — Identifying cost reduction opportunities in AI spend
- Product Managers — Building LLM-powered features with budget constraints
- DevOps/Platform Engineers — Implementing caching and batching infrastructure
Best For
- Analyzing API billing data to identify cost drivers and inefficiencies
- Redesigning prompts for token efficiency without losing quality
- Evaluating trade-offs between model tiers and task-specific performance
- Planning infrastructure changes (caching, batching, queue optimization)
- Quantifying ROI before implementing cost-reduction changes
You might also like
You can audit current staffing models across all group practice locations, identify cost-saving opportunities and scheduling inefficiencies, analyze compensation competitiveness against market benchmarks, and model the financial impact of proposed staffing changes. This skill helps you make data-driven staffing decisions that balance patient demand, clinician utilization, payroll efficiency, and staff retention across your multi-clinic operation.
by aymericzip
You can conduct comprehensive sports injury evaluations that integrate mechanism of injury analysis, special orthopedic tests, and clinical reasoning to determine injury severity and return-to-sport readiness. This skill helps you create sport-specific treatment plans, document findings for third-party clearance, and differentiate between structural damage, functional limitations, and performance deficits—enabling you to make evidence-based decisions about athlete management and injury prevention strategies.
by aymericzip
Conduct comprehensive fall risk assessments for patients 65+ by identifying modifiable risk factors including balance deficits, medication effects, neurological changes, and environmental hazards. Generate customized intervention plans that adjust treatment protocols based on individual risk profiles, with integrated documentation for liability protection and clinical decision-making that prioritizes patient safety throughout the care episode.
by aymericzip
You systematically evaluate sports injuries using clinical orthopedic testing procedures, document findings within the context of sport-specific biomechanical demands, and generate individualized rehabilitation protocols with clear functional milestones. This skill helps you create athlete-centered treatment plans that satisfy insurance documentation requirements while establishing realistic return-to-play timelines that athletes, coaches, and parents can understand and commit to.
by aymericzip
You can systematically coordinate complex schedules across multiple clinical disciplines, eliminate bottlenecks in patient flow, allocate shared rooms and equipment efficiently, and establish clear handoff procedures between providers. This skill helps you resolve scheduling conflicts, reduce provider idle time and patient wait times, and design communication protocols that keep all disciplines aligned on patient status and treatment plans.
by aymericzip
You can conduct thorough fall risk assessments for geriatric patients by integrating validated screening instruments, functional mobility tests, spinal and neurological evaluation, and environmental risk factor analysis. The skill guides you through systematic documentation of findings, identifies modifiable risk factors specific to chiropractic intervention, and generates personalized prevention protocols that reduce fall probability while preserving patient autonomy and quality of life.
by aymericzip
You can process multimedia content across multiple formats: transcribe audio files up to 9.5 hours with timestamps and speaker identification, analyze images for objects, text, and visual relationships, extract structured data from PDF documents including tables and charts, and generate images from text prompts with refinement capabilities. This skill unifies fragmented multimodal workflows into one coherent Claude-powered interface, eliminating context-switching between different tools.
by trungtv
You can build AI-powered Python agents quickly using the @agentic decorator for simple functions or spawn() for full multi-agent systems with behavioral control. Agents maintain persistent state, integrate with MCP tools, handle typed return values, and coordinate with each other through a unified framework—enabling everything from basic AI-enhanced functions to complex agent orchestration pipelines.
by aymericzip