AGENT SKILL

twinmind-cost-tuning

Optimize TwinMind costs across Free, Pro, and Enterprise tiers. Use when analyzing usage patterns, reducing costs, or choosing the right pricing tier for your needs. Trigger with phrases like "twinmind cost", "reduce twinmind spending", "twinmind pricing optimization", "twinmind budget".

License 许可
MIT
最近更新
2026-11-15
Stars
★ 2.0k
Forks
233

01 前言 Front Matter

name
twinmind-cost-tuning
author
Jeremy Longshore <jeremy@intentsolutions.io>
license
MIT
version
1.0.0
description
Optimize TwinMind costs across Free, Pro, and Enterprise tiers. Use when analyzing usage patterns, reducing costs, or choosing the right pricing tier for your needs. Trigger with phrases like "twinmind cost", "reduce twinmind spending", "twinmind pricing optimization", "twinmind budget".
allowed-tools
Read, Write, Edit

03 SKILL.md 正文

TwinMind Cost Tuning

Overview

Optimize TwinMind costs through usage analysis, tier selection, and efficiency improvements.

Prerequisites

  • TwinMind account with billing access
  • Understanding of usage patterns
  • Access to usage analytics

Instructions

Step 1: Understand Pricing Structure

// src/twinmind/costs/pricing.ts
export interface TierPricing {
  name: string;
  monthlyBase: number;          // Base subscription cost
  transcriptionRate: number;    // Per hour of audio
  apiRequestRate: number;       // Per 1000 requests
  aiTokenRate: number;          // Per 1M tokens
  storageRate: number;          // Per GB per month
  includedHours: number;        // Included transcription hours
  includedTokens: number;       // Included AI tokens
}

export const pricingTiers: Record<string, TierPricing> = {
  free: {
    name: 'Free',
    monthlyBase: 0,
    transcriptionRate: 0,        // Unlimited but basic quality
    apiRequestRate: 0,           // No API access
    aiTokenRate: 0,
    storageRate: 0,
    includedHours: Infinity,     // Unlimited
    includedTokens: 500000,
  },
  pro: {
    name: 'Pro',
    monthlyBase: 10,
    transcriptionRate: 0.23,     // $0.23 per hour (Ear-3)
    apiRequestRate: 0,           // Included
    aiTokenRate: 0,              // Included up to limit
    storageRate: 0,
    includedHours: Infinity,     // Unlimited
    includedTokens: 2000000,
  },
  enterprise: {
    name: 'Enterprise',
    monthlyBase: 0,              // Custom pricing
    transcriptionRate: 0.15,     // Volume discount
    apiRequestRate: 0,
    aiTokenRate: 0,
    storageRate: 0,
    includedHours: Infinity,
    includedTokens: Infinity,
  },
};

Step 2: Analyze Current Usage

// src/twinmind/costs/analyzer.ts
export interface UsageData {
  period: string;
  transcriptionHours: number;
  apiRequests: number;
  aiTokensUsed: number;
  storageGB: number;
  meetings: number;
}

export interface CostAnalysis {
  currentTier: string;
  currentCost: number;
  estimatedCostPerTier: Record<string, number>;
  recommendedTier: string;
  potentialSavings: number;
  breakdown: CostBreakdown;
}

export interface CostBreakdown {
  base: number;
  transcription: number;
  api: number;
  aiTokens: number;
  storage: number;
  total: number;
}

export async function analyzeUsage(): Promise<CostAnalysis> {
  const client = getTwinMindClient();

  // Get last 30 days usage
  const usage = await client.get('/usage', {
    params: { period: 'last_30_days' },
  });

  const data: UsageData = usage.data;
  const currentTier = (await client.get('/account')).data.plan;

  // Calculate cost for each tier
  const costByTier: Record<string, number> = {};

  for (const [tierName, tier] of Object.entries(pricingTiers)) {
    const cost = calculateTierCost(tier, data);
    costByTier[tierName] = cost;
  }

  // Find recommended tier
  const validTiers = Object.entries(costByTier)
    .filter(([tier]) => meetsRequirements(tier, data))
    .sort(([, a], [, b]) => a - b);

  const recommendedTier = validTiers[0]?.[0] || 'pro';
  const currentCost = costByTier[currentTier];
  const recommendedCost = costByTier[recommendedTier];

  return {
    currentTier,
    currentCost,
    estimatedCostPerTier: costByTier,
    recommendedTier,
    potentialSavings: currentCost - recommendedCost,
    breakdown: calculateBreakdown(pricingTiers[currentTier], data),
  };
}

function calculateTierCost(tier: TierPricing, usage: UsageData): number {
  let cost = tier.monthlyBase;

  // Transcription cost (if over included)
  if (usage.transcriptionHours > tier.includedHours) {
    cost += (usage.transcriptionHours - tier.includedHours) * tier.transcriptionRate;
  }

  // AI tokens (if over included)
  if (usage.aiTokensUsed > tier.includedTokens) {
    const overageTokens = usage.aiTokensUsed - tier.includedTokens;
    cost += (overageTokens / 1000000) * tier.aiTokenRate;
  }

  return cost;
}

function meetsRequirements(tier: string, usage: UsageData): boolean {
  // Free tier doesn't have API access
  if (tier === 'free' && usage.apiRequests > 0) {
    return false;
  }
  return true;
}

Step 3: Implement Cost Monitoring

// src/twinmind/costs/monitor.ts
export interface CostAlert {
  type: 'warning' | 'critical';
  message: string;
  currentSpend: number;
  threshold: number;
  percentUsed: number;
}

export interface BudgetConfig {
  monthlyBudget: number;
  warningThreshold: number;  // e.g., 0.8 for 80%
  criticalThreshold: number; // e.g., 0.95 for 95%
  notifications: {
    email?: string[];
    slack?: string;
    webhook?: string;
  };
}

export class CostMonitor {
  private config: BudgetConfig;

  constructor(config: BudgetConfig) {
    this.config = config;
  }

  async checkBudget(): Promise<CostAlert | null> {
    const client = getTwinMindClient();
    const usage = await client.get('/usage/cost', {
      params: { period: 'current_month' },
    });

    const currentSpend = usage.data.total_cost;
    const percentUsed = currentSpend / this.config.monthlyBudget;

    if (percentUsed >= this.config.criticalThreshold) {
      return {
        type: 'critical',
        message: `TwinMind spending at ${(percentUsed * 100).toFixed(1)}% of monthly budget`,
        currentSpend,
        threshold: this.config.monthlyBudget,
        percentUsed,
      };
    }

    if (percentUsed >= this.config.warningThreshold) {
      return {
        type: 'warning',
        message: `TwinMind spending at ${(percentUsed * 100).toFixed(1)}% of monthly budget`,
        currentSpend,
        threshold: this.config.monthlyBudget,
        percentUsed,
      };
    }

    return null;
  }

  async sendAlert(alert: CostAlert): Promise<void> {
    const { notifications } = this.config;

    if (notifications.slack) {
      await sendSlackNotification(notifications.slack, {
        text: `:${alert.type === 'critical' ? 'rotating_light' : 'warning'}: ${alert.message}`,
        blocks: [
          {
            type: 'section',
            text: {
              type: 'mrkdwn',
              text: `*TwinMind Budget Alert*\n${alert.message}\n\nCurrent: $${alert.currentSpend.toFixed(2)} / $${alert.threshold.toFixed(2)}`,
            },
          },
        ],
      });
    }

    if (notifications.email) {
      await sendEmail({
        to: notifications.email,
        subject: `[${alert.type.toUpperCase()}] TwinMind Budget Alert`,
        body: `${alert.message}\n\nCurrent spend: $${alert.currentSpend.toFixed(2)}\nBudget: $${alert.threshold.toFixed(2)}\nUsed: ${(alert.percentUsed * 100).toFixed(1)}%`,
      });
    }
  }
}

// Schedule budget checks
export function startCostMonitoring(config: BudgetConfig): void {
  const monitor = new CostMonitor(config);

  // Check every hour
  setInterval(async () => {
    const alert = await monitor.checkBudget();
    if (alert) {
      await monitor.sendAlert(alert);
    }
  }, 60 * 60 * 1000);
}

Step 4: Optimize Token Usage

// src/twinmind/costs/token-optimization.ts
export interface TokenUsageReport {
  totalTokens: number;
  byFeature: Record<string, number>;
  optimizationPotential: number;
  recommendations: string[];
}

export async function analyzeTokenUsage(): Promise<TokenUsageReport> {
  const client = getTwinMindClient();
  const usage = await client.get('/usage/tokens', {
    params: { period: 'last_30_days', breakdown: 'by_feature' },
  });

  const byFeature = usage.data.breakdown;
  const totalTokens = Object.values(byFeature).reduce((sum: number, v) => sum + (v as number), 0);

  const recommendations: string[] = [];
  let optimizationPotential = 0;

  // Check for optimization opportunities
  if (byFeature.summary > totalTokens * 0.5) {
    recommendations.push('High summary token usage - consider using shorter summaries');
    optimizationPotential += byFeature.summary * 0.3;
  }

  if (byFeature.chat > totalTokens * 0.3) {
    recommendations.push('High chat token usage - implement response caching');
    optimizationPotential += byFeature.chat * 0.2;
  }

  if (byFeature.memory_search > totalTokens * 0.2) {
    recommendations.push('Consider indexing optimization for memory searches');
    optimizationPotential += byFeature.memory_search * 0.1;
  }

  return {
    totalTokens,
    byFeature,
    optimizationPotential,
    recommendations,
  };
}

// Token-efficient summary options
export const tokenEfficientOptions = {
  summary: {
    brief: {
      maxLength: 150,          // Short summary
      includeActionItems: true,
      includeKeyPoints: false,
      estimatedTokens: 200,
    },
    standard: {
      maxLength: 300,
      includeActionItems: true,
      includeKeyPoints: true,
      estimatedTokens: 500,
    },
    detailed: {
      maxLength: 500,
      includeActionItems: true,
      includeKeyPoints: true,
      estimatedTokens: 800,
    },
  },
};

Step 5: Implement Usage Quotas

// src/twinmind/costs/quotas.ts
export interface QuotaConfig {
  dailyTranscriptionHours: number;
  dailyApiRequests: number;
  dailyAiTokens: number;
}

export class QuotaManager {
  private config: QuotaConfig;
  private usage = {
    transcriptionHours: 0,
    apiRequests: 0,
    aiTokens: 0,
    lastReset: new Date(),
  };

  constructor(config: QuotaConfig) {
    this.config = config;
    this.startDailyReset();
  }

  private startDailyReset(): void {
    // Reset at midnight
    const now = new Date();
    const midnight = new Date(now);
    midnight.setHours(24, 0, 0, 0);

    const msUntilMidnight = midnight.getTime() - now.getTime();

    setTimeout(() => {
      this.resetUsage();
      setInterval(() => this.resetUsage(), 24 * 60 * 60 * 1000);
    }, msUntilMidnight);
  }

  private resetUsage(): void {
    this.usage = {
      transcriptionHours: 0,
      apiRequests: 0,
      aiTokens: 0,
      lastReset: new Date(),
    };
  }

  canTranscribe(hours: number): boolean {
    return this.usage.transcriptionHours + hours <= this.config.dailyTranscriptionHours;
  }

  canMakeApiRequest(): boolean {
    return this.usage.apiRequests < this.config.dailyApiRequests;
  }

  canUseTokens(tokens: number): boolean {
    return this.usage.aiTokens + tokens <= this.config.dailyAiTokens;
  }

  recordUsage(type: 'transcription' | 'api' | 'tokens', amount: number): void {
    switch (type) {
      case 'transcription':
        this.usage.transcriptionHours += amount;
        break;
      case 'api':
        this.usage.apiRequests += amount;
        break;
      case 'tokens':
        this.usage.aiTokens += amount;
        break;
    }
  }

  getRemainingQuota(): {
    transcriptionHours: number;
    apiRequests: number;
    aiTokens: number;
  } {
    return {
      transcriptionHours: Math.max(0, this.config.dailyTranscriptionHours - this.usage.transcriptionHours),
      apiRequests: Math.max(0, this.config.dailyApiRequests - this.usage.apiRequests),
      aiTokens: Math.max(0, this.config.dailyAiTokens - this.usage.aiTokens),
    };
  }
}

// Usage with middleware
export function quotaMiddleware(quotaManager: QuotaManager) {
  return (req: Request, res: Response, next: NextFunction) => {
    if (!quotaManager.canMakeApiRequest()) {
      return res.status(429).json({
        error: 'Daily API quota exceeded',
        remaining: quotaManager.getRemainingQuota(),
      });
    }

    quotaManager.recordUsage('api', 1);
    next();
  };
}

Output

  • Pricing structure analysis
  • Usage cost analyzer
  • Budget monitoring system
  • Token optimization strategies
  • Quota management

Cost Optimization Strategies

Strategy Savings Implementation
Use brief summaries 30-50% tokens Set maxLength: 150
Cache memory searches 20-30% Implement result caching
Batch transcriptions 10-15% Group small files
Off-peak processing 0% (no TwinMind variation) N/A
Annual billing 33% Switch to annual plan

Pricing Comparison

Feature Free Pro ($10/mo) Enterprise
Transcription Unlimited Unlimited Unlimited
API Access No Yes Yes
Ear-3 Model No Yes Yes
Context Tokens 500K 2M Unlimited
Rate Limits 30/min 60/min 300/min
Support Community 24hr Dedicated
Cost/Hour $0 $0.23 Custom

Error Handling

Issue Cause Solution
Unexpected charges Usage spike Set up budget alerts
Quota exceeded High usage Implement rate limiting
Wrong tier Poor planning Analyze usage patterns

Resources

Next Steps

For reference architecture, see twinmind-reference-architecture.

04 同仓库更多 Skills

Dicklesworthstone
Dicklesworthstone/pi_agent_rust

theme-factory

Toolkit for styling artifacts with a theme. These artifacts can be slides, docs, reportings, HTML landing pages, etc. There are 10 pre-set themes with colors/fonts that you can apply to any artifact that has been creating, or can generate a new theme on-the-fly.

视觉设计
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

web-artifacts-builder

Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.

前端开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

webapp-testing

Toolkit for interacting with and testing local web applications using Playwright. Supports verifying frontend functionality, debugging UI behavior, capturing browser screenshots, and viewing browser logs.

测试与质量前端开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

neuropixels-analysis

Neuropixels neural recording analysis. Load SpikeGLX/OpenEphys data, preprocess, motion correction, Kilosort4 spike sorting, quality metrics, Allen/IBL curation, AI-assisted visual analysis, for Neuropixels 1.0/2.0 extracellular electrophysiology. Use when working with neural recordings, spike sorting, extracellular electrophysiology, or when the user mentions Neuropixels, SpikeGLX, Open Ephys, Kilosort, quality metrics, or unit curation.

★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

skill-creator

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends Claude's capabilities with specialized knowledge, workflows, or tool integrations.

Skill与智能体开发
★ 2.0k
Dicklesworthstone
Dicklesworthstone/pi_agent_rust

notebooklm

Use this skill to query your Google NotebookLM notebooks directly from Claude Code for source-grounded, citation-backed answers from Gemini. Browser automation, library management, persistent auth. Drastically reduced hallucinations through document-only responses.

Skill与智能体开发
★ 2.0k