Chiang Mai YouTubers community

Launch a Faceless YouTube Channel (Hands-On)

This is the agenda for the workshop. Keep this page open during the workshop; we’ll open links directly from here.

Join the Chiang Mai YouTubers Facebook Group →

Register for the Workshop

Save your spot for Saturday, September 27, 2025 — 2:00–4:00 PM at Weave Artisan Society (FREE)

Register now →

Workshop companion

Faceless Channel: Planning & Best AI Tools

  • Host Green Light Studio — Stephen, Cliffton, Jake, Zeke (with Craig)
  • When Sat, Sept 27, 2025 • 2:00–4:00 PM (2 hours)
  • Venue Weave Artisan Society, Chiang Mai
  • Price FREE
  • Format Hands-on walkthrough • Open laptops encouraged • Doors close at 2:30 PM
  • Audience Beginners & intermediate creators, editors, and entrepreneurs
Faceless YouTube workflow overview: planning to AI-assisted production

Green Light Studio

Chiang Mai–based marketing & media team. Services include video editing, podcasts, YouTube channel management, and marketing coaching. They focus on clear plans, sustainable execution, and measurable results.

Visit site

Weave Artisan Society

A creative campus in Chiang Mai hosting this workshop. Weave is launching a new coworking space — follow their page for opening details, events, and membership info.

Follow Weave Events

What you’ll learn (outcomes)

  • How to plan faceless content for consistency: cadence, topics, and formats that suit voiceover-first production.
  • Which AI tools fit each stage: script, images, video, music, and voice — with a simple way to compare costs and limits.
  • A repeatable folder + naming system that supports automation and teams.
  • How to design a fast first cut flow: prompts → beats → temp cut → polish.
  • Leave with a 30‑day action plan and a small set of templates you can re‑use.

Agenda (real timeline, 2:00–4:00 PM)

Doors close at 2:30 PM. Keep this page open during the session.

2:00–2:30
KickoffIntroductions & soft opening

Light social time while everyone arrives and opens the companion page. Doors close at 2:30 PM.

  • What a faceless channel optimizes for: repeatable cadence, recognizable visual system, and safe/licensed assets.
  • How the workshop flows: quick demos inside popular web UIs → simple, repeatable pipelines you can copy 1:1.
  • Baseline folder grammar used all day: /research, /script, /images, /video_raw, /video_final, /music, /voice, /project, /licenses, /logs.
2:30–2:40
PlanFaceless channel content planning

We set the production target first so tools serve the format, not the other way around.

  • Pick a cadence (e.g., 1–2/week) and show format (essay, listicle, explainer, biography).
  • Define 5–7 core beats (Hook → Context → Beat 1–3 → Wrap/CTA). Aim for 120–160 wpm voiceover pacing.
  • File naming for sanity and search: YYYY-MM-DD_topic_step_index.ext (e.g., 2025-09-27_riddleofrome_script_01.md).
2:40–2:52
WriteBest AI tools for script writing (web UIs first)

We’ll stay inside friendly web apps most creators already use, compare popular choices by strength, and show how to get clean, reusable outputs.

Most-used web UIs (fastest to adopt)
  • ChatGPT (OpenAI): excellent general writing, outline → draft in one pass, strong editing (rewrite/tone). Great for hooks and punch-ups.
  • Claude (Anthropic): very strong long-form structure and “explain like I’m 5” clarity; good at research synthesis and safer outputs.
  • Gemini (Google): good grounding with Google services, fast ideation; solid for bullet→paragraph expansion.
  • Mistral Chat (Mistral): concise style, quick iteration; helpful for beat scaffolds and shorter YouTube formats.
  • Notion AI: integrated in notes; convenient for turning bullet research into a first draft inside your workspace.
What we actually do live
  • Paste tagged research → use a “prompt pack” to output a beat map and a VO-ready draft (plain paragraphs, no markdown fluff).
  • Save two files: /script/beats.json (title, hook, beats, CTA) and /script/vo_draft.md (clean narration).
  • Tip: keep the exact prompt and model name in /logs/script_YYYY-MM-DD.json for reproducibility.
Quick pick by style
  • Explainers/essays: Claude or ChatGPT.
  • Listicles/shorts: ChatGPT or Mistral Chat.
  • Research-heavy: Claude or Gemini (for structured notes then draft).
2:52–3:04
VisualsBest AI tools for image generation (thumbnails & on-screen)

We compare popular web generators creators actually use, then show how to keep outputs consistent with your brand system.

Most-used web UIs
  • Midjourney: aesthetic consistency and composition control; great for striking thumbnails. Closed platform (no general public API); export and file-name carefully.
  • Adobe Firefly: integrated with Photoshop/Express; strong text rendering and brand-friendly compositing for thumbnail titles.
  • Leonardo.ai: versatile presets and texture control; decent batch runs for scene packs.
  • Ideogram: strong typography inside generated images; useful for word-heavy thumbnails.
  • Playground / Canva Magic Media: accessible for quick iterations and team collaboration.
What we actually do live
  • Generate 2–3 thumbnail candidates at 1280×720 (16:9) with large, high-contrast focal points and minimal text.
  • Export naming: thumb-A_seed1234_1280x720.png, insert-map_1x1_seed4321.png; keep a tiny .json sidecar with prompt/seed.
  • On-screen inserts: generate simple, brand-colored visuals (maps, silhouettes, objects) that read at mobile sizes.
Quick pick by need
  • Bold art style: Midjourney or Leonardo.
  • Text-in-image: Ideogram or Firefly.
  • Team-friendly edits: Firefly or Canva.
3:04–3:16
MotionBest AI tools for video generation (b-roll & transitions)

Use web generators for short b-roll clips and transitions, then assemble a clean first cut. We focus on what’s actually usable for faceless channels.

Most-used web UIs
  • Runway Gen-3: high-quality motion for stylized b-roll, transitions, and abstract backgrounds.
  • Pika: quick text/image→video with playful motion; great for short inserts and looping elements.
  • Luma Dream Machine: strong photoreal tendencies; good for cinematic textures and moody b-roll.
  • Kaiber / CapCut AI: accessible motion stylization and quick cuts for short-form assets.
What we actually do live
  • Plan 3–5 insert shots from the beat map (e.g., “factory exterior,” “map zoom,” “abstract data flow”).
  • Render short 3–6 s clips at your project fps (23.976 or 29.97). Keep a swap-friendly timeline: each insert on its own track, easy to replace.
  • Normalize before editing: consistent resolution/codec, and no variable frame rates; store clips under /video_raw.
Quick pick by need
  • Stylized transitions: Runway or Pika.
  • Photoreal textures: Luma.
  • Simplest workflow: CapCut AI (then export “highest” and normalize).
3:16–3:24
SoundBest AI tools for music generation (beds, stingers, stems)

Get simple, safe beds that support narration. We’ll compare popular song-level tools and show how to keep levels consistent.

Most-used web UIs
  • Suno & Udio: fast text→music; great for cinematic beds and stingers; export WAV/MP3 and stems where available.
  • Soundraw / AIVA / Beatoven: library-like feel with structure controls; useful for background loops.
  • Boomy / Mubert: quick loops and genre pads; handy for shorts and repetitive cues.
What we actually do live
  • Create 2 beds at different energy levels (low/medium), 1–2 short stingers; export 48 kHz WAV.
  • Keep dialogue-first mixing: beds around −23 to −18 LUFS under VO; avoid masking frequencies (dip 2–4 kHz).
  • Name for reuse: bed_100bpm_minor_low.wav, stinger_rise_02.wav.
Quick pick by need
  • Cinematic pads: Suno / Udio.
  • Edit-friendly loops: Soundraw / AIVA.
  • Shorts/quick loops: Mubert / Boomy.
3:24–3:28
VoiceBest AI tools for VoiceOver (web UIs)

Reliable narration is the backbone of faceless channels. We’ll compare the most popular TTS tools creators actually use.

Most-used web UIs
  • ElevenLabs: natural prosody, multi-style voices, strong cloning; easy batch rendering.
  • Play.ht: clear narration voices, SSML controls, good stability for long reads.
  • Descript Overdub: perfect for script-edit ↔ voice fix loops; integrated text-based editing.
  • Azure Neural TTS / Google Cloud TTS / Amazon Polly (via consoles): reliable, many languages; good for standardized VO.
What we actually do live
  • Export 48 kHz mono WAV per paragraph; keep a pronunciation list (names, terms) for consistency.
  • Target loudness for VO around −16 LUFS; add 200–300 ms room tone head/tail for cleaner edits.
Quick pick by need
  • Most natural narration: ElevenLabs.
  • Fast pipeline edits: Descript Overdub.
  • Enterprise consistency: Azure / Google / Polly.
3:28–3:30
OptionalAPIs & self-hosting (quick overview)

Two-minute tour of how to automate later. We won’t build it live, but here’s the map.

  • APIs: most tools above offer REST/SDKs; you can save prompts/settings and auto-export to your folders.
  • Self-host: Stable Diffusion/SDXL for images, Stable Video Diffusion/AnimateDiff for motion, Coqui XTTS/Piper/VITS for TTS; run locally when you need privacy/cost control.
  • Rule of thumb: start in web UIs until your format stabilizes, then automate the boring parts.
3:30–4:00
CommunityOpen discussion & social

Compare setups, trade prompts/presets, and pair up for accountability. We’ll stick around to help you lock installs and templates.

  • Swap prompt packs (hooks, beats), thumbnail style notes, VO settings, and music bed recipes.
  • Optional partner plan: write two concrete deliverables for next week and exchange contact details.