lindy-performance-tuning
Optimize Lindy AI agent performance and response times. Use when improving latency, optimizing throughput, or reducing response times. Trigger with phrases like "lindy performance", "lindy slow", "optimize lindy", "lindy latency".
- License 许可
- MIT
- 最近更新
- 2026-11-15
- Stars
- ★ 2.0k
- Forks
- 233
01 前言 Front Matter
- name
- lindy-performance-tuning
- author
- Jeremy Longshore <jeremy@intentsolutions.io>
- license
- MIT
- version
- 1.0.0
- description
- Optimize Lindy AI agent performance and response times. Use when improving latency, optimizing throughput, or reducing response times. Trigger with phrases like "lindy performance", "lindy slow", "optimize lindy", "lindy latency".
- allowed-tools
- Read, Write, Edit
03 SKILL.md 正文
Lindy Performance Tuning
Overview
Optimize Lindy AI agent performance for faster response times and higher throughput.
Prerequisites
- Production Lindy integration
- Baseline performance metrics
- Access to monitoring tools
Instructions
Step 1: Measure Baseline Performance
import { Lindy } from '@lindy-ai/sdk';
interface PerformanceMetrics {
avgLatency: number;
p95Latency: number;
p99Latency: number;
throughput: number;
errorRate: number;
}
async function measureBaseline(agentId: string, iterations = 100): Promise<PerformanceMetrics> {
const lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });
const latencies: number[] = [];
let errors = 0;
const start = Date.now();
for (let i = 0; i < iterations; i++) {
const runStart = Date.now();
try {
await lindy.agents.run(agentId, { input: 'Benchmark test' });
latencies.push(Date.now() - runStart);
} catch (e) {
errors++;
}
}
const totalTime = (Date.now() - start) / 1000;
latencies.sort((a, b) => a - b);
return {
avgLatency: latencies.reduce((a, b) => a + b, 0) / latencies.length,
p95Latency: latencies[Math.floor(latencies.length * 0.95)],
p99Latency: latencies[Math.floor(latencies.length * 0.99)],
throughput: iterations / totalTime,
errorRate: errors / iterations,
};
}
Step 2: Optimize Agent Instructions
// BEFORE: Verbose instructions (slow)
const slowAgent = {
instructions: `
You are a helpful assistant that should carefully consider
each request and provide detailed, comprehensive responses.
Think step by step about each query. Consider all possibilities.
Provide examples and explanations for everything.
`,
};
// AFTER: Concise instructions (fast)
const fastAgent = {
instructions: `
Be concise. Answer directly. Skip pleasantries.
Format: [Answer] (1-2 sentences)
`,
config: {
maxTokens: 100, // Limit response length
},
};
Step 3: Enable Streaming
// Non-streaming (waits for full response)
const result = await lindy.agents.run(agentId, { input });
console.log(result.output); // Logs after full response
// Streaming (immediate partial responses)
const stream = await lindy.agents.runStream(agentId, { input });
for await (const chunk of stream) {
process.stdout.write(chunk.delta); // Real-time output
}
Step 4: Implement Caching
import NodeCache from 'node-cache';
const cache = new NodeCache({ stdTTL: 300 }); // 5 minute TTL
async function runWithCache(agentId: string, input: string) {
const cacheKey = `${agentId}:${crypto.createHash('md5').update(input).digest('hex')}`;
// Check cache
const cached = cache.get(cacheKey);
if (cached) {
return cached;
}
// Run agent
const lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });
const result = await lindy.agents.run(agentId, { input });
// Cache result
cache.set(cacheKey, result);
return result;
}
Step 5: Optimize Concurrency
// Poor: Sequential execution
for (const input of inputs) {
await lindy.agents.run(agentId, { input }); // Slow!
}
// Better: Parallel with controlled concurrency
import pLimit from 'p-limit';
const limit = pLimit(5); // Max 5 concurrent
const results = await Promise.all(
inputs.map(input =>
limit(() => lindy.agents.run(agentId, { input }))
)
);
Performance Checklist
[ ] Baseline metrics captured
[ ] Instructions are concise
[ ] Max tokens configured appropriately
[ ] Streaming enabled for long responses
[ ] Caching implemented for repeated queries
[ ] Concurrency optimized
[ ] Connection pooling enabled
[ ] Timeout values tuned
Output
- Baseline performance metrics
- Optimized agent configuration
- Caching implementation
- Concurrency patterns
Error Handling
| Issue | Cause | Solution |
|---|---|---|
| High latency | Verbose instructions | Simplify prompts |
| Timeouts | Response too long | Add max tokens |
| Throttling | Too concurrent | Limit parallelism |
Examples
Complete Performance Client
class PerformantLindyClient {
private lindy: Lindy;
private cache: NodeCache;
private limiter: any;
constructor() {
this.lindy = new Lindy({ apiKey: process.env.LINDY_API_KEY });
this.cache = new NodeCache({ stdTTL: 300 });
this.limiter = pLimit(5);
}
async run(agentId: string, input: string) {
const cacheKey = `${agentId}:${input}`;
const cached = this.cache.get(cacheKey);
if (cached) return cached;
const result = await this.limiter(() =>
this.lindy.agents.run(agentId, { input })
);
this.cache.set(cacheKey, result);
return result;
}
}
Resources
Next Steps
Proceed to lindy-cost-tuning for cost optimization.
04 同仓库更多 Skills
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.
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.
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.
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.
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.
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.