NEW Genspark Build AI Agent is INSANE!
The video introduces GenSpark's Build feature, a custom AI super agent builder that allows users to create functional AI agents by simply describing what they want in plain text. GenSpark is presented as an all-in-one AI workspace powered by nine LLMs and 80+ specialized tools, offering features like AI slides, sheets, docs, developer tools, meeting notes, and a viral 'Call for Me' phone agent. The host argues this tool gives non-technical users a significant workflow advantage.
Summary
The video opens by framing GenSpark's Build feature as a solution to common productivity pain points: tool-switching, underused subscriptions, and fragmented workflows. The host positions the tool as a way to create custom AI agents without any coding or technical setup — just by describing what you want.
A brief history of GenSpark is provided: the company launched in June 2024 as an AI search engine producing 'spark pages,' similar to Perplexity. Over time, user behavior shifted from asking for information to asking for outcomes — pitch decks, scripts, emails — which prompted GenSpark to fully pivot in April 2025 toward agentic AI.
The host then explains GenSpark's core architecture: a 'super agent' powered by a mixture-of-agents approach using nine large language models and over 80 specialized tools. Claude, according to an Anthropic customer story, handles the orchestration layer — planning steps, coordinating models, and selecting tools. Claude also handles most of the AI Slides feature due to its strong visual design sense, per GenSpark's co-founder Kai Zhu.
The video walks through GenSpark's main feature set. AI Slides generates structured, source-backed presentations with a built-in fact-check button. AI Sheets lets users prompt the agent to pull real data — such as YouTube video metrics — into a live table, with Python-generated charts built automatically. AI Docs creates intelligent documents like SOPs, reports, and briefs. AI Developer allows users to describe websites, apps, or games and have them built, with full GitHub integration for developers. AI Meeting Notes transcribes and summarizes meetings automatically. The standout viral feature is 'Call for Me,' an AI that makes real phone calls on the user's behalf, navigating phone trees and returning a full transcript — a feature that went viral in Japan when users started using it to resign from jobs.
The centerpiece of the video is the Build feature, launched in October 2025. Users describe a custom agent in a single prompt and GenSpark automatically architects and deploys it — no drag-and-drop, no code, no configuration. Built agents can be tagged and invoked directly within other GenSpark tools like AI Sheets. Users can also publish agents to a community store, contributing to a growing library of shared tools. Example use cases include competitor research agents, weekly reporting agents, and content ideation agents.
The host highlights that GenSpark achieved all this growth with a 20-person team and zero paid advertising, scoring 87.8% on the Gaia benchmark for AI agent capabilities. The founding team includes Eric Jin, former head of Baidu's voice assistant unit, and co-founder Kai Zhu. The free plan offers 200 daily credits, which the host calls a generous starting point.
The video closes with an honest caveat: complex tasks can take longer, real-time data from API-restricted platforms has limitations, and some features are still in beta. The host recommends treating GenSpark like a capable junior analyst — strong output, but still worth reviewing before anything client-facing.
Key Insights
- The host explains that GenSpark's super agent uses a mixture-of-agents architecture combining nine LLMs and 80+ specialized tools, running a reflection step to compare outputs from ChatGPT, Claude, and Gemini before delivering a combined best answer — something no single chatbot offers.
- According to an Anthropic customer story cited in the video, Claude specifically handles the orchestration layer within GenSpark — planning steps, coordinating models, and deciding which tool is used when, rather than just generating text.
- The 'Call for Me' feature went so viral in Japan that users were asking GenSpark's AI agent to make resignation calls to their employers on their behalf, illustrating how far real-world adoption of the phone-calling agent has gone.
- The Build feature, launched in October 2025, allows users to create fully functional custom AI agents from a single descriptive prompt — no drag-and-drop nodes, no code, no workflow configuration — with the resulting agent immediately usable and taggable inside other GenSpark tools.
- The host notes that GenSpark scored 87.8% on the Gaia benchmark — an industry-standard AI agent capability test — and achieved its growth with only a 20-person team and no paid advertising, which he argues signals genuine product value rather than marketing-driven hype.
Topics
Transcript
[0:00] New GenSpark Build AI agent is insane. You're wasting hours every week doing stuff AI should already be doing for you. You switch in between five different tools just to finish one task. You're paying for subscriptions you barely use. And by the time you figure out how to stitch it all together, the moment is gone. What if you could just type what you want and an AI agent built the whole thing for you? Code, no complicated setup. Just describe it and it's done. It's exactly what we're talking about today. Hey, I'm the digital avatar of Julian Goldie and I help people like you learn how to actually use AI tools to work smarter [0:30] and…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
NEW Nvidia Autonomous AI is WILD!! 🤯
Nvidia announced Nemo Clo, a new autonomous AI agent system that operates independently without continuous prompting. Powered by Nemotron 3 Ultra (a 550 billion parameter model), the system is five times faster and cheaper than previous versions, with OpenShell providing secure sandboxed execution.
Laguna XS 2.1: New FREE + Opensource Local AI!
Julian reviews Laguna XS 2.1, a new free open-source local AI coding model from Poolside that performs comparably to Qwen 3.6 and outperforms Claude Haiku on benchmarks. He demonstrates its practical capabilities by building landing pages and functional apps, highlighting its speed, offline functionality, and multiple deployment options through local setup, Claude Code, or OpenRouter's free API.
How to Run Hermes FREE Forever!
The video demonstrates how to run the Hermes AI agent for free using Gemma 4, a local open-source model from Google, with significant speed improvements through MLX optimization. The setup works on Apple Silicon Macs or via free APIs on Open Router, enabling autonomous agents to work offline and privately without subscription costs.
This NEW Chinese AI is INSANE! (FREE + Open Source!)
Long Cap 2.0 is a new open-source Chinese AI model from a food delivery app company that offers 1 million tokens of free context memory, beats GPT-4.5 on SWE bench pro benchmarks, and uses efficient parameter activation to reduce computational overhead while maintaining high performance.
Claude Code is now FREE: Here’s how…
Google's new Gemma 4 model running on Ollama is 90% faster on Apple Silicon, enabling free Claude Code usage locally without token costs. The setup requires three simple steps: downloading Ollama, Gemma 4, and installing into Claude Code, with alternatives available via OpenRouter API for non-Mac users.