Create Professional Videos Using Hermes Agent + Remotion (2026 Tutorial)

Summary

The video demonstrates an AI-assisted workflow for creating a polished vertical video using Hermes Agent and Remotion, without manually editing in tools such as Premiere Pro or DaVinci Resolve. The example video explains a claim about methane detected on interstellar comet 3I/Atlas.

Workflow

1. Set up the project

• Create a new Remotion project in an empty folder.
• Choose the blank-project option and install Tailwind CSS.
• Install the Remotion skills, selecting the skills to use and Hermes Agent as the target agent.
• Keep the installation scoped to the project, then start Hermes in that folder.

2. Research a topic

Ask Hermes to find ideas for a short-form video and research them online. Have it produce a document containing multiple ideas, sources, and a recommendation. The presenter chooses the idea about the James Webb Space Telescope detecting methane on interstellar comet 3I/Atlas.

3. Write the script and plan

Ask Hermes to draft a voiceover script, narration guidance, storyboard cues, and production notes. Then generate a scene-by-scene production plan that includes details such as runtime, canvas, frame rate, visual style, color palette, and the intended look of each scene.

4. Generate visual assets

Configure an image-generation model in Hermes Desktop under Capabilities → Tools → Image Generation. Use the configured provider to create images for the video.

5. Build and review the video

Ask Hermes to create the video from the production plan using the installed Remotion skills. Hermes renders the scenes and checks selected frames against the plan, making changes when they do not match. The first render is a motion-graphics video without audio.

6. Add audio

Configure a text-to-speech model under Capabilities → Tools → Text-to-Speech. The presenter notes that local or paid providers can be used. Have Hermes generate the voiceover and add background music and sound effects using Remotion’s sound packs or user-provided audio assets.

7. Add animated captions

Prompt Hermes to create captions suited to short-form platforms. The requested style includes accurate word-level timestamps, emphasis on the currently spoken word, and subtle pop or scale transitions rather than static text. Hermes uses a Whisper model to obtain word-level timing information and apply it to the captions.

8. Revise and reuse

Make further changes by describing them clearly in prompts. You can also add custom visual or audio assets and specify a brand color palette. The presenter says the same workflow can be adapted for long-form videos and product advertisements.

Example Video Content

The generated narration presents methane as a chemical clue about the comet’s formation environment, not as evidence of alien life. It describes methane, water, and carbon dioxide around the comet’s nucleus and contrasts the comet’s composition with that of ordinary Solar System comets. The subtitles show a repeated line at the end of the voiceover.

Speakers and Sources Featured

• Presenter: The tutorial’s narrator, who explains and demonstrates the workflow.
• AI-generated voiceover: Narrates the example video about comet 3I/Atlas.
• Tools and models named: Hermes Agent, Remotion, GPT Image 2, and Whisper.
• Scientific source mentioned: The James Webb Space Telescope. No specific research paper or web source is named in the subtitles.