May 3
Loading video player...
If you’re building AI agents with tools like OpenClaw, LangChain, or Claude Code, there’s a good chance you’re massively overpaying for LLM usage without realizing it. In this video, we see how I reduced my AI agent costs by 70% with one simple change—no rewrites, no complex routing logic, just smarter model selection. I’ll show you exactly how Manifest, an open-source LLM router, sits between your agent and 600+ models to automatically route each request to the cheapest capable option in under 2 milliseconds. 🔗 Relevant Links Manifest Repo - https://github.com/mnfst/manifest Manifest Site - https://manifest.build/ ❤️ More about us Radically better observability stack: https://betterstack.com/ Written tutorials: https://betterstack.com/community/ Example projects: https://github.com/BetterStackHQ 📱 Socials Twitter: https://twitter.com/betterstackhq Instagram: https://www.instagram.com/betterstackhq/ TikTok: https://www.tiktok.com/@betterstack LinkedIn: https://www.linkedin.com/company/betterstack 📌 Chapters: 0:00 Reduce AI Agent Costs (70% Savings Explained) 0:37 Why AI Agents Are So Expensive (Hidden LLM Costs) 0:50 The Real Problem: Bad LLM Routing 1:45 Live Demo: 70% Cheaper AI Agent (Manifest Setup) 2:57 What Manifest Is (Open-Source LLM Router Explained) 3:18 Why Most LLM Calls Don’t Need GPT-4 / Claude Opus 3:53 Manifest vs OpenRouter vs LiteLLM (Best LLM Router?) 4:50 Self-Hosted vs Cloud LLM Routing (Privacy + Cost) 5:09 Pros & Cons of Manifest (Dev Feedback) 6:20 Is Manifest Worth It? (Who Should Use It) 6:45 Final Verdict: Cut Your AI Costs Without Rewriting Code