BEARTH SIGNAL
Daily AI Briefing for Digital Nomads
0
Episodes
0/8
Today's Steps
0
Day Streak 🔥
Episode Pipeline
Quick Actions
📋 Step 1 — Copy this prompt into Perplexity AI
Step 2 — Paste Perplexity Report
0 characters — aim for 500+
Initializing...
Titles
Script
Shorts
Twitter
LinkedIn
Instagram
TikTok
Pinterest
Clips
YouTube titles, thumbnail brief, description, and chapter markers will appear here.
🖼 YouTube Thumbnail — FLUX.1 · 576×1024 (9:16)
Your YouTube video script will appear here after generation.
3 standalone 45–60 second Shorts/Reels scripts will appear here.
5 Twitter/X posts will appear here.
Your LinkedIn post will appear here.
Your Instagram caption and 30 hashtags will appear here.
📸 Instagram Post Image — FLUX.1-schnell · 512×640 (4:5)
Auto-filled from episode topic after generation. Edit before generating.
Your TikTok caption will appear here.
Your Pinterest pin description will appear here.
📌 Pinterest Pin Image — FLUX.1-schnell · 512×768 (2:3)
Auto-filled from episode topic after generation. Edit before generating.
5 clip extraction briefs (which moments to cut from the full script).
⚡ Performance (recommended): Optimized for Genspark Voice Agent — voice directions converted to natural pacing (pauses, emphasis, rhythm). Produces dynamic, non-monotone speech.
Full: Keeps raw emotional tags [Lean in, whisper] — for ElevenLabs or advanced XTTS-v2.
Clean: Pure spoken words, zero markup — fallback for basic TTS engines.
Full: Keeps raw emotional tags [Lean in, whisper] — for ElevenLabs or advanced XTTS-v2.
Clean: Pure spoken words, zero markup — fallback for basic TTS engines.
TTS-ready script will appear here after generating the YouTube script.
Production markers [VISUAL:], [CHAPTER:], [START/END VIRAL CLIP] are stripped automatically.
⚡ Performance mode translates voice directions into natural pacing cues (pauses, emphasis, CAPS) that your Genspark Voice Agent interprets as vocal dynamics — creating realistic highs, lows, whispers, and energy shifts.
Production markers [VISUAL:], [CHAPTER:], [START/END VIRAL CLIP] are stripped automatically.
⚡ Performance mode translates voice directions into natural pacing cues (pauses, emphasis, CAPS) that your Genspark Voice Agent interprets as vocal dynamics — creating realistic highs, lows, whispers, and energy shifts.
Step 3 — Generate Voice Audio
Copy the ⚡ Performance script above, then paste into your Genspark Voice Agent:
Workflow: Copy "⚡ Performance" script → paste into your Genspark Voice Agent (Saimen's voice is stored there) → generate → download MP3
Next: MP3 + Saimen photo → HunyuanVideo-Avatar → Video
Do NOT use generic audio tools — the Genspark Agent has the exact voice profile stored.
Next: MP3 + Saimen photo → HunyuanVideo-Avatar → Video
Do NOT use generic audio tools — the Genspark Agent has the exact voice profile stored.
Production Stack
One tap to open each tool in the BEARTH SIGNAL workflow.
STEP 1 — RESEARCH
Perplexity AI
Real-time AI research. Use the exact prompt below — it's engineered to return structured data perfect for BEARTH SIGNAL scripts.
STEP 2 — GENERATE ALL CONTENT
BEARTH SIGNAL Generator
Paste your research report → generate 9 content pieces in ~60 seconds. Titles, full script, 3 Shorts scripts, and all 5 social platform posts.
Cost: ~$0.06/day using Claude Sonnet (script) + Haiku (social). ~$1.80/month total.
STEP 3 — AI VOICE AUDIO
Voice Generation (Genspark Agent)
Go to the 🎙 Audio tab for TTS-ready script (auto-cleaned). Copy "⚡ Performance" version → open your Genspark BEARTH SIGNAL Voice Agent (has Saimen's voice stored) → paste script → generate.
⚠️ IMPORTANT: Use your custom Genspark Voice Agent — it has Saimen's voice already stored. Do NOT use other audio tools.
🎙 Audio tab →
🎙 Audio tab →
⚡ Performance → paste into Genspark Voice Agent → generate → download MP3VOICE TOOLS
XTTS-v2 / MegaTTS 3
XTTS-v2 clones Saimen's voice from a 30s reference. MegaTTS 3 is simplest (no code, browser-only). Both free via HuggingFace.
STEP 4 — AI AVATAR VIDEO
HunyuanVideo-Avatar (HuggingFace — Free)
Generate the Saimen Bearth AI avatar video. Upload a photo + the XTTS-v2 audio → full-body lip-synced video. Free alternative to HeyGen ($29+/mo). Setup: docs/huggingface-stack.md
Also needed: XTTS-v2 for voice (huggingface.co/coqui/XTTS-v2) + MuseTalk for lip sync. Full workflow in docs/huggingface-stack.md
STEP 5 — PUBLISH
YouTube Studio
Upload the rendered video. Use the generated titles, description, and tags from the "Titles" tab. Enable "Contains AI-generated content" checkbox.
Recommended upload time: 9:00 AM UTC daily (peak nomad audience window)
STEP 6 — AUTO-POSTING
Make.com
Automated social distribution. One webhook trigger from this app sends all 5 social pieces (Twitter, LinkedIn, Instagram, TikTok, Pinterest) to Make.com for scheduling.
See
docs/make-setup.md for webhook configuration.Episode Workflow
Progress: 0 / 8 done
Archive
All generated episodes. Tap to reload any episode's content.