I Gave Claude 7 Days To Make the Most Realistic Murder Mystery Game (Fish Audio)

I Gave Claude 7 Days To Make the Most Realistic Murder Mystery Game (Fish Audio)

Title history
  • I Gave Claude 7 Days To Make the Most Realistic Murder Mystery Game (Fish Audio) — observed on 20261101-0208 UTC (YouTube)

by Pat Simmons · View the channel on PSI

Applied Sciences & Engineering Computing & Cybersecurity UMComputer programmingUMKGames development and programming
🎙 Pat Simmons 👥 30K 📅 October 31, 2026 ⏱ 33 min 👁 9K 📄 tutorial 🧭 2026-11-01
Available in: English (current) Français

Keywords

Claudegame developmentAI agentsvoice synthesisBlender

Summary

In this video, Pat Simmons demonstrates how he built a fully playable murder mystery game entirely using a team of Claude Opus 5.5 agents over seven days, without writing a single line of code himself. The game features 3D environments created in Blender, realistic character voices generated with Fish Audio’s Drama 3 model, and a complete storyline developed by an AI writers’ room. Simmons walks through every step: story generation, voice casting, character art, mesh and texture creation, and finally coding the game to run in the browser. He also breaks down the costs, totaling about $4,200 in API equivalent pricing (largely from 3D generation), with voices costing only about $1.42 and GPU rentals at $25. He emphasizes that his process is not optimized, and offers a “Game Studio” skill on his blog to help others build similar games more efficiently. The video includes many examples of failed iterations and technical challenges, highlighting the iterative nature of working with AI agents. It concludes with lessons learned, such as using headless Blender via Python scripts and leveraging open-source models like FlashHead for realistic mouth movements.

184 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides high practical value by offering a step-by-step walkthrough of building a game with AI agents, including specific tools, costs, and failures. The argumentation is solid as it includes live demonstrations of the game, cost breakdowns, and explanations of each step. However, the content is based on a single personal experiment and may not generalize, and some claims about model capabilities are not independently verified.

Scientific Rigor, Source Quality, Title Accuracy

The video references several tools and services (Fish Audio, GPT Image Gen, FlashHead, RunPod) and provides links to his blog for the skill and game. The title accurately describes the content, and the video fully delivers on the promise of a 7-day build of a murder mystery game. There are no formal citations, but the video is transparent about the costs and limitations. The sponsor mention is clearly disclosed. The video adheres to the title.

157 words

Title / Content Match

Title accurately reflects the content: building a murder mystery game with Claude agents over 7 days, sponsored by Fish Audio.

Quality & Reliability

7/10

Video provides transparent breakdown of costs and process, but relies on personal experience and unverified models.

Chapters

Cited Sources

Contribution & Novelties

This video offers a unique, hands-on account of using a team of AI agents to create a complete game from scratch, including realistic 3D graphics and voice acting. It provides practical insights into cost optimization and iterative development with AI. The open-source Game Studio skill is a novel contribution that others can reuse.

Pour aller plus loin :

102 words

Radar Profile

The radar profile shows balanced scores across quality and quantity, with slightly lower reliability due to the informal nature of the content. The technical level is moderate, indicating the video is accessible to intermediate users.

Reliability 6/10