AI Video Editor is a course in directing AI agents such as Codex CLI and Claude Code to run a video-editing pipeline with Hyperframes: removing silences with FFmpeg, transcribing Thai with Whisper, and adding kinetic captions and a brand kit. It suits business owners, content creators and marketing teams, and you leave able to render several aspect ratios from one source.
Producing video content every day is one of the most time-consuming tasks for business owners and marketing teams today. Editing a single clip with click-and-drag tools (CapCut, Premiere Pro) takes 1-3 hours, including cutting silences, adding subtitles, finding B-roll and exporting multiple aspect ratios for each platform. When you must produce a clip a day, it becomes a bottleneck that stalls business growth. This course shifts the mindset from editing by hand to instructing an AI agent to run the pipeline for you, using professional tools that run on the learner's own machine: Codex CLI (OpenAI) for scripting and running the pipeline end-to-end, Claude Code (Anthropic) for tasks that require understanding file structure and deep template customization,
and Hyperframes (HeyGen, Apache 2.0), an open-source video framework that uses HTML/CSS as its source, so you can change captions, motion and brand colors by editing just a few lines of code, with no per-clip cost. A key feature is that learners install and test the entire system on their own machine in class, testing with real footage they bring, so by the end of the day they have a working pipeline ready to use the next day and take home a real, finished first clip. The course is designed so those who have never used an AI agent can learn from scratch, though they should be comfortable with basic command line use. (1 day, 6 hours, small groups of 12 per session.)
Course Details
A 1-day course, 6 hours (09:30-16:30), delivered as a hands-on workshop where learners set up the system on their own machine. Small groups of 12 per session, beginner-friendly (start from scratch). Learners bring their own notebook and a 5-10 minute sample footage. By the end of the day they take home an AI-agent video editing pipeline, a brand kit, and a real finished first clip in all 3 aspect ratios.
No public rounds are open right now. Join the waiting list and we will contact you first when the next round opens, or ask us on LINE. Or call 02-570-8449 or 088-807-9770
Run FFmpeg's silencedetect filter over the audio, setting a threshold for what counts as silence, such as -30dB, and a minimum duration for a silent gap. FFmpeg reports the start and end time of each silence, so you can cut out only the parts with speech as short clips and then join them back into a single file. With an AI agent such as Claude Code or Codex CLI, you can ask it to write a script that handles every step.
Whisper comes in several sizes, from tiny, base and small to medium and large. Larger models transcribe Thai more accurately but need more time and memory, so choose one that balances your machine's power with the speed you need, and always correct typos, names and technical terms. Output can be exported as SRT or JSON with word-level timestamps, which you can turn into kinetic captions that highlight each word.
Separate the content from the layout: keep the footage, captions and logo in one composition, then set the frame size and element positions for each aspect ratio, such as 1080x1920 for Reels and TikTok or 1920x1080 for YouTube. Keep the subject inside the safe area for every size, and place captions where app buttons will not cover them. This course practises rendering all three ratios from a single source with Hyperframes.
It suits business owners who produce video regularly, content creators, digital marketing teams and video editors who want to speed up with automation. You do not need to code, because the AI agent writes the code, but you should be confident with a computer, know basic command-line use (reviewed in class) and understand video basics such as aspect ratio and frame rate.
You will be able to install Codex CLI, Claude Code and Hyperframes with Node.js, FFmpeg and Whisper on your own machine, run a pipeline that removes silences, turns Thai speech into kinetic captions, applies a brand kit and pulls B-roll from the Pexels API, and render several aspect ratios from one source. You will also prompt AI agents effectively and plan batch production. Participants who complete the training receive a Certificate of Completion with its own number, which can be verified online.
Bring your own notebook computer, because you will install and configure everything on it, including Node.js, FFmpeg, Whisper, Codex CLI, Claude Code and Hyperframes, and store API keys securely. You should also bring 5 to 10 minutes of sample footage to edit into a real clip during the class.
Articles are published in Thai.
E-commerce and online-store businesses juggle many things at once in a highly competitive market: writing product details and titles so they get found, creating product images and ad media, running marketing and campaigns, answering customer chats and…
Marketing today demands fast content across many channels, constantly adjusted to data. Many marketing teams face the same problems: too much work, not enough time for creative thinking, and repetitive tasks such as writing captions, drafting emails,…
Customer service work involves a large volume of messages and repeated questions: answering chats, emails, comments and social messages, handling complaints and dissatisfied customers, finding the right information to reply accurately, and keeping the…
