Grammarly Performance Tuning

Latency Benchmarks

Instructions

Cache Score Results

import { LRUCache } from 'lru-cache';
import { createHash } from 'crypto';

const scoreCache = new LRUCache<string, any>({ max: 500, ttl: 3600000 });

async function cachedScore(text: string, token: string) {
  const key = createHash('sha256').update(text).digest('hex');
  const cached = scoreCache.get(key);
  if (cached) return cached;
  const score = await grammarlyClient.score(text);
  scoreCache.set(key, score);
  return score;
}

Parallel API Calls

// Score + AI detect in parallel (they're independent)
async function fullAudit(text: string, token: string) {
  const [score, ai] = await Promise.all([
    grammarlyClient.score(text),
    grammarlyClient.detectAI(text),
  ]);
  return { score, ai };
}

Resources

• Grammarly API

Next Steps

For cost optimization, see grammarly-cost-tuning.