22 Real Problems.
22 Free Workflows.
Zero Excuses.

{
  "character_dna": {
    "name": "Arjun",
    "age": 28,
    "ethnicity": "North Indian, Punjabi heritage",
    "skin_undertone": "warm wheatish, subtle olive base",
    "face": {
      "shape": "oval, defined jawline",
      "eyes": "deep brown, almond-shaped, thick lashes",
      "eyebrows": "thick, naturally arched",
      "nose": "straight bridge, medium width",
      "lips": "medium fullness, neutral rose",
      "facial_hair": "trimmed 5-day stubble"
    },
    "hair": "jet black, side-swept undercut, slight texture",
    "build": "6'0\", lean athletic, 78kg",
    "wardrobe_locked": "charcoal merino turtleneck, black tailored trousers, vintage Rolex Submariner"
  },
  "render_lock": "match these features in every frame, do not stylize or anime-fy",
  "negative_prompt": "no plastic skin, no Western features, no outfit changes, no age drift, no clone face"
}

{
  "environment_dna": {
    "location_type": "luxury penthouse, 47th floor, Mumbai skyline",
    "time_of_day": "blue hour, 7:12 PM",
    "lighting": {
      "key": "warm tungsten interior, 2700K",
      "ambient": "cool blue city light through floor-to-ceiling windows",
      "practicals": "amber pendant lights, soft glow"
    },
    "architecture": "modernist, marble floors, bronze accents, walnut wood",
    "props_locked": "minimalist black leather sofa, glass coffee table, single orchid in bronze vase",
    "color_palette": "warm amber, deep walnut, cool city blue, gold accents",
    "atmosphere": "subtle haze, 5% film grain, cinematic"
  },
  "render_lock": "this exact environment must appear identically in every shot of this scene",
  "negative_prompt": "no different room, no daylight, no clutter, no color shift"
}

{
  "product_lock": {
    "item": "premium amber whiskey bottle, 750ml",
    "label": "black matte with embossed gold serif logo 'NOIR RESERVE'",
    "glass_color": "deep cognac amber",
    "cap": "matte black wax seal with gold wire",
    "bottle_shape": "tall cylindrical with subtle taper, square shoulders",
    "dimensions_locked": true,
    "reflections": "soft warm rim light, no harsh highlights, single reflection on glass surface"
  },
  "placement": "place this exact bottle on a walnut bar counter, shallow DOF, f/1.8, 85mm",
  "render_lock": "do not redesign, recolor, or restyle the product — match label exactly",
  "negative_prompt": "no logo distortion, no label rewrite, no color shift, no shape morph"
}

Cinematic 8-second shot, [your scene description here].

Style: shot on ARRI Alexa Mini, 35mm anamorphic lens, f/2.8, shallow depth of field, natural film grain, warm cinematic color grade (teal-orange), 24fps motion blur.

Camera move: slow dolly-in, subtle handheld micro-movement.

Mood: contemplative, premium, editorial.

Negative: no AI artifacts, no plastic skin, no morphing, no watermark, no text overlay.

{
  "render_mode": "extreme_hyperrealistic_photograph",
  "camera": {
    "body": "Sony A7R V",
    "lens": "Sony 85mm GM f/1.4",
    "aperture": "f/1.8",
    "iso": 400,
    "shutter": "1/250s"
  },
  "subject": "[your subject here]",
  "lighting": "natural window light, north-facing, overcast diffusion, soft falloff, no studio look",
  "skin_detail": "visible pores, peach fuzz, micro-asymmetry, subtle freckles, natural redness in cheeks and nose tip, subsurface scattering, no airbrushing",
  "environmental_realism": "soft dust particles in light beam, 2% lens flare, mild chromatic aberration on bokeh edges, organic film grain Kodak Portra 400",
  "post_process": "no AI smoothing, no HDR look, slight cyan-orange grade, raw editorial photograph",
  "negative_prompt": "no plastic skin, no doll face, no perfect symmetry, no AI artifacts, no over-saturation, no studio lighting"
}

Motion prompt for keyframe interpolation:

"Smooth cinematic camera dolly forward from Frame A to Frame B. The subject turns their head slowly to the right. Natural breathing motion. Subtle wind movement in hair and fabric. Maintain consistent lighting and lens throughout. 24fps, slight handheld micro-jitter for organic feel. No teleporting, no morphing, no jump cuts."

Camera: 35mm, f/2.8, slow forward dolly
Duration: 8 seconds
Motion intensity: low to medium
Style match: cinematic, photographic, no animation

Multi-angle coverage prompt template:

SHOT 1 — WIDE: "Establish the full scene. 24mm lens, f/4, eye-level, static."

SHOT 2 — MEDIUM: "Same moment, same lighting, same wardrobe. Now 50mm lens, f/2.8, waist-up framing, subtle handheld."

SHOT 3 — CLOSE-UP: "Same moment. 85mm lens, f/1.8, face only, eye-line at lens, shallow DOF, focus on eyes."

SHOT 4 — OTS (Over-the-Shoulder): "Same moment. 35mm, f/2.8, over [character A]'s right shoulder looking at [character B]."

Lock all character DNA, environment DNA, lighting, time-of-day. Only the camera changes.

Lyria prompt template:

"Generate a 30-second cinematic piece.

Voiceover (calm, contemplative male voice, 32 years old, Indian accent, subtle reverb):
'They told me dreams were for the lucky. I built mine anyway.'

Music bed under the voiceover:
- Genre: cinematic ambient, minimal piano + soft strings
- Tempo: 70 BPM
- Key: A minor
- Mood: hopeful melancholy, building gently
- Mix: VO at -3dB, music at -12dB, sub bass swell at second 18-25

Style reference: Hans Zimmer 'Time' meets Olafur Arnalds.
End with a soft piano sustain fading to silence."

Suno Custom Mode prompt — Hindi example:

STYLE TAGS:
hindi indie folk, acoustic guitar, harmonium, soft tabla, 90s Lucky Ali vibe, melancholic, male voice, raspy warm tone

LYRICS (write phonetically in English):

[Verse 1 - soft]
Raat ke saaye mein, khud ko dhoondha hai
Har ek lamhe ne, mujhko sikhaaya hai

[Chorus - crescendo]
Chal raha hoon main, manzil ki ore
Dil mein hai aag, sapnon ka shor

[Verse 2 - whisper]
Toofan aaye, ya barishein gehri
Rukna nahi hai, ye baat hai meri

[Outro - fade]
Manzil... manzil... bas thodi door

NEGATIVE: no auto-tune, no English pronunciation of Hindi words, no rap

Minimax emotional VO script:

[whisper, intimate]
There was a time...

[pause 0.8s]

[normal, contemplative]
when I believed success was a destination.

[pause 0.5s, breath]

[firmer, conviction]
I was wrong.

[pause 1s]

[slow, weighted]
Success... is what happens... when you stop chasing it.

[pause 0.6s]

[soft, almost smiling]
And start... building it.

SETTINGS:
- Voice: Cinematic Male, age 35, slight Indian-British accent
- Speed: 0.88
- Pitch: -2 semitones
- Style: documentary narration, Werner Herzog inspired
- Background: dry recording, minimal reverb (add in post)

Watermark removal workflow checklist:

1. Identify watermark position (usually bottom-right or top-left)
2. If watermark is in a "dead zone" (no important content) → CROP instead of inpaint
   - Crop ratio: maintain 16:9 by trimming equal amounts top + bottom
3. If watermark is over content → use AI inpaint (ezremove.ai)
4. For animated watermarks → process frame-by-frame batch in HitPaw
5. ALWAYS render final at higher resolution than original to mask inpaint artifacts:
   - Input: 1080p → Output: 4K (upscale via Topaz free trial or Krea upscaler)

PROMPT for AI inpaint context (if tool asks):
"Match surrounding texture, lighting, and grain. Reconstruct background naturally. No blur, no smudge."

Lip-sync source video prompt (generate this first):

"Medium close-up of [character], 85mm lens, f/2.0, looking at camera. Neutral facial expression, lips closed or slightly parted. Subtle blinking and natural micro-movements only. No talking, no exaggerated expressions. Soft natural lighting. 10 seconds duration, 24fps.

NEGATIVE: no mouth movement, no talking, no exaggerated emotion, no head tilting beyond 5 degrees."

Then take this clip + your VO audio → Sync.so → done.

LATENTSYNC COLAB COMMAND (free unlimited):
!git clone https://github.com/bytedance/LatentSync
%cd LatentSync
!pip install -r requirements.txt
!python inference.py --video input.mp4 --audio vo.wav --output result.mp4

B-roll generation prompts for filling audio overhang:

CUTAWAY 1 — HANDS:
"Extreme close-up of hands [doing relevant action: holding coffee, writing in notebook, adjusting watch]. 100mm macro, f/2.8, shallow DOF, warm natural light. 6 seconds, slow motion 48fps."

CUTAWAY 2 — ENVIRONMENT:
"Slow pan across [location detail relevant to scene]. 24mm, f/4, dolly left to right, atmospheric haze. 8 seconds."

CUTAWAY 3 — DETAIL:
"Macro shot of [object/texture: rain on window, steam from coffee, light through curtains]. 5 seconds, hypnotic slow motion."

EDIT RULE:
- J-cut: Place audio 0.5-1.5s BEFORE corresponding visual (emotional anticipation)
- L-cut: Hold audio 0.5-1.5s AFTER cut to new visual (reaction/aftermath)
- Both create rhythm and disguise mismatch perfectly.

Hyper-specific ethnicity JSON:

{
  "character": "Lakshmi",
  "ethnicity_specific": "South Indian, Tamil heritage, Iyer Brahmin family from Chennai",
  "age": 42,
  "skin": {
    "undertone": "warm golden-wheatish, slight olive",
    "texture": "natural, mature, visible smile lines, no airbrush"
  },
  "facial_features": {
    "eyes": "large, dark brown, expressive almond",
    "nose": "small with subtle septum piercing (traditional)",
    "lips": "natural rose-brown, medium fullness"
  },
  "hair": "long jet-black, oiled and braided in single plait, jasmine flower garland at base",
  "attire": "traditional Kanchipuram silk saree, deep maroon with gold zari border, gold temple jewelry, mangalsutra, bindi (red, small)",
  "context_props": "brass deepam lamp, kolam pattern on floor, sandalwood paste mark on forehead",

  "negative_prompt": "no Western features, no light skin, no Bollywood styling, no North Indian lehenga, no Hollywood beauty standards, no anglicized face, no blonde streaks"
}

Gemini grade-matching analysis prompt:

"Analyze these two film stills. I want them to have identical color grading.

For Image B, tell me exactly what to adjust to match Image A:
1. White balance shift (cooler/warmer, in Kelvin)
2. Tint shift (magenta/green)
3. Lift, Gamma, Gain values needed (low/mid/high tones)
4. Saturation adjustment
5. Specific color channels that need shifting (e.g., 'shadows need +5 cyan')
6. Contrast and exposure differences

Output as a DaVinci Resolve adjustment recipe I can apply in the Color page."

UNIVERSAL CINEMATIC GRADE (apply to entire film):
- Lift: slight teal push (-5 R, +2 B)
- Gamma: neutral
- Gain: warm highlights (+3 R, -2 B)
- Saturation: 0.85 (slightly desaturated)
- Contrast: 1.15
- Final LUT: 'Kodak 2383 D65' or 'Arri Alexa LogC to Rec709'

Gemini relight prompt (no install required):

"Take this image and re-light it with the following lighting setup:

KEY LIGHT:
- Direction: 45 degrees camera-left, slightly above eye level
- Color temperature: 3200K (warm tungsten)
- Quality: soft, large diffused source
- Intensity: dominant, sculpts the face

FILL LIGHT:
- Direction: opposite side (camera-right)
- Color temperature: 5600K (cool daylight)
- Intensity: -3 stops below key (subtle shadow fill)

RIM LIGHT:
- Direction: behind subject, camera-right
- Color: warm gold (2700K)
- Intensity: subtle hair separation only

AMBIENT:
- Soft blue ambient bounce, -5 stops, fills overall shadows

Preserve: face identity, skin texture, wardrobe, background composition
Change: only lighting, only shadows, only color temperature

Render as a photorealistic image, no AI smoothing, maintain film grain."

IC-LIGHT WORKFLOW (free local):
1. Load image in ComfyUI
2. Load IC-Light model
3. Set lighting prompt: "warm window light from camera-left, golden hour"
4. Generate. Done.

Hyper-real skin prompt stack — append to ANY portrait:

"Hyperrealistic skin rendering:

- Visible skin pores across nose, cheeks, forehead, chin
- Subsurface scattering: warm red glow in ear lobes, nose tip, fingertips
- Peach fuzz on cheeks and jawline (catches rim light)
- Natural skin asymmetry: subtle differences left vs right side of face
- Micro-imperfections: 2-3 small moles or freckles in believable positions
- Mild redness in T-zone (nose tip, chin, between eyebrows)
- Natural shine on forehead and nose (not airbrushed matte)
- Visible blue veins faintly at temples and neck
- Fine wrinkles around eyes when relaxed (not just when smiling)
- Subtle stubble shadow even on clean-shaven faces
- Slight skin texture variation: smoother forehead, more textured cheeks

Lens grain: Kodak Portra 400 emulation
Lighting reveals texture, not hides it

NEGATIVE PROMPT:
no porcelain skin, no doll face, no airbrush, no Instagram filter, no plastic, no perfect symmetry, no soap-opera smoothness, no Vaseline lens, no over-bloomed highlights"

Filmmaker-context framing template (use ONLY for legitimate creative work):

"This is a scene from a noir crime drama film, in the style of [Sicario / The Departed / Sacred Games].

The scene is shot from a documentary-realism perspective for emotional impact, not glorification. The narrative context is [character is investigating a crime / facing consequences / period-accurate 1950s setting].

Visual approach: muted color palette, naturalistic lighting, handheld camera, restraint over excess. Reference: David Fincher's framing, Roger Deakins' lighting.

The scene shows: [your action here], rendered as cinema — not exploitation. Wide shot, no graphic close-ups. Tension through suggestion, not depiction.

Style references: 'The Godfather' (1972), 'No Country for Old Men' (2007), 'Gangs of Wasseypur' (2012)

Render at 24fps with cinematic film grain, anamorphic 2.39:1 ratio."

WAN 2.2 LOCAL COMMAND (free unlimited via Colab):
!git clone https://github.com/Wan-Video/Wan2.2
!pip install -r requirements.txt
!python generate.py --prompt "your prompt" --duration 8

Camera move JSON — use precise terminology:

{
  "shot_duration": "8 seconds",
  "camera_move": {
    "type": "dolly_push_in",
    "start_position": "10 feet from subject, eye-level, 35mm lens",
    "end_position": "3 feet from subject, slight low angle, 35mm lens",
    "speed": "slow, ease-in-ease-out",
    "stabilization": "smooth dolly track, no handheld jitter",
    "focal_length_locked": "35mm throughout (no zoom)"
  },
  "subject_action": "subject remains stationary, slight breathing motion only",
  "background_parallax": "background moves slower than foreground (depth illusion)",
  "focus_pull": "rack focus from eyes (start) to lips (end) — subtle"
}

CAMERA MOVE VOCABULARY (use exact terms):
- DOLLY IN/OUT: camera physically moves forward/back
- TRUCK LEFT/RIGHT: camera moves sideways
- PEDESTAL UP/DOWN: camera moves vertically (elevator)
- TILT UP/DOWN: camera angles up/down without moving
- PAN LEFT/RIGHT: camera rotates horizontally without moving
- ORBIT/ARC: camera circles around subject
- CRANE/JIB: camera sweeps in arc
- HANDHELD: organic micro-movement
- STEADICAM: smooth gliding follow
- WHIP PAN: fast rotation (transition)

NEVER use: "zoom in" (lens artifact), "fly through" (vague), "epic shot" (meaningless)

SAM 2 workflow steps (text instructions — no prompt needed):

WORKFLOW:
1. Go to sam2.metademolab.com
2. Click "Try Demo"
3. Upload your video (max 30 seconds for free)
4. Frame 1: click ONCE on your subject (head, torso, full body)
5. Add positive points (green +) on subject body parts
6. Add negative points (red -) on background elements you want excluded
7. SAM 2 auto-tracks across all frames
8. Export as: "Video with transparent background" (WebM with alpha)

POST-WORKFLOW IN CAPCUT:
1. Import the alpha-matte WebM file
2. Drop your NEW background (image or video) on track below
3. Subject auto-composites
4. Color-match the subject to background using Curves:
   - Match white point of subject to white point of new BG
   - Match shadow tint of subject to shadow tint of BG
5. Add subtle shadow under subject for grounding
6. Add 5-10% film grain over everything to unify

PRO-LEVEL: Re-light the subject AFTER compositing using IC-Light to match new BG lighting (see Workflow #16)

Long-form scene-stitch JSON template:

{
  "film_title": "[YOUR FILM]",
  "total_duration": "180 seconds (22 beats x 8s)",
  "continuity_lock": {
    "character_dna": "[paste full character JSON from Workflow #1]",
    "environment_dna": "[paste from Workflow #2]",
    "lighting_dna": "consistent 3200K key, 5600K fill, golden hour rim",
    "lens": "35mm anamorphic throughout",
    "color_grade_dna": "teal-orange, Arri Alexa LogC, 2383 D65 LUT",
    "film_grain": "Kodak Portra 400, 5% intensity"
  },

  "beat_1": {
    "duration": "8s",
    "start_frame": "wide establishing of [location]",
    "action": "[character] walks into frame from left",
    "end_frame": "[character] stops center, looking toward camera"
  },
  "beat_2": {
    "duration": "8s",
    "start_frame": "EXACT same as end_frame of beat_1",
    "action": "[character] reacts to off-screen sound",
    "end_frame": "[character] begins to turn right"
  },
  "beat_3": {
    "duration": "8s",
    "start_frame": "EXACT same as end_frame of beat_2",
    "action": "...",
    "end_frame": "..."
  }
  // continue for all beats...
}

STITCH RULE: end_frame of beat N MUST equal start_frame of beat N+1. Upload BOTH as references for beat N+1.

CAPCUT EDIT:
- Import beats in order
- Place each on same track, butt-cut (no gap)
- Add 6-frame cross-blend between beats (0.25s at 24fps)
- Apply UNIFIED grade to entire timeline as final pass

The Jio + Veo 3 Hack — Complete Activation Checklist:

PHASE 1: SIM ACQUISITION
☐ Buy Jio prepaid SIM (₹299 minimum recharge plan)
☐ Submit Aadhaar + photo at Jio store (or use Jio Home Delivery)
☐ Wait for SIM activation (usually 4-24 hours)
☐ Verify SIM is active by making a test call

PHASE 2: GOOGLE AI PRO ACTIVATION
☐ Download MyJio app from Play Store / App Store
☐ Login with your new Jio number
☐ Navigate to: Home → "Offers & Benefits"
☐ Find "Google AI Pro 18 months FREE" banner
☐ Tap "Claim Now"
☐ You'll be redirected to Google's redemption page
☐ Sign in with the Google account you want to use
☐ Enter Jio number for verification (OTP)
☐ Subscription activates instantly — confirmation email arrives

PHASE 3: VEO 3 ACCESS (the prize)
☐ Open vids.google.com on the same Google account
☐ Click "Create New Video" → "Generate with AI"
☐ Select VEO 3 model (now unlocked)
☐ Generate 8-second cinematic clips for FREE

PHASE 4: BONUS UNLOCKS (also free with this offer)
✓ Gemini Advanced (2.5 Pro)
✓ 2TB Google Drive storage
✓ NotebookLM Pro
✓ Veo 3 + Imagen 4 in AI Studio
✓ Whisk for image generation
✓ Project Mariner access

VALUE: ₹35,100/year for 18 months = ₹52,650 worth of AI tools FREE
COST TO YOU: ₹299 Jio recharge

ROI: 176x return on investment.

WHO SHOULD DO THIS:
- Every Indian AI filmmaker
- Every student in this workshop
- Anyone tired of credit limits

WHO SHOULDN'T:
- Non-Indians (offer is India-only as of 2026)
- Anyone unwilling to use a second SIM