Extract & Transcribe 1000 Reels

Project Description
I have a dataset of roughly one thousand Instagram Reels that follow a consistent structure. I need each Reel broken down into profile images and a structured text transcription of the conversation.
Each Reel typically:
Starts with 3 profile photos of a girl
Then shows a chat conversation (dating app style)
Requirements
1. Image Extraction
Extract and save the 3 profile images shown at the beginning of each Reel
Images should be clean (no UI overlays, buttons, or cropping issues)
2. Conversation Transcription
Transcribe the full conversation shown in the Reel
Maintain the exact order of messages
Format strictly as:

Man: "message"
GirlName: "message"
App: "message"
Use:
“Man” for the male speaker
The girl’s actual name when available
“App” for any system-generated messages (e.g., notifications, prompts)
Do not include timestamps
Ignore filler sounds unless they affect meaning
Focus only on the visible/readable conversation text (not audio narration)
3. File Structure & Naming
Each Reel should produce 4 files:
GirlName_p01.jpg → Profile image 1
GirlName_p02.jpg → Profile image 2
GirlName_p03.jpg → Profile image 3
GirlName_tx.txt → Transcribed conversation
Example:

Sophia_p01.jpg
Sophia_p02.jpg
Sophia_p03.jpg
Sophia_tx.txt
4. Organization
Files can be delivered grouped per Reel or in batches
Naming consistency is critical
5. Notes
Accuracy and consistency are more important than speed
You may use automation tools (scraping, OCR, etc.), but final output must be clean and error-free
Please specify:
Your approach (manual vs automated)
Estimated turnaround time for ~1000 Reels
Price per Reel or per batch