CD & Vinyl Product Data Enrichment

Project overview
We are an online retailer of CDs and Vinyl LPs. Our supplier data sheets are missing key information we need to sell products. We need a full data sweep of an excel sheet containing 60,000 lines. A further sheet of approximately 180,000 lines will follow for the right freelancer or team.

What we need for each product:

1. Descriptions (HTML)
One paragraph about the artist or band, and one paragraph about this specific release.
Concise and factual, using verified information only. No invented promotional claims or unsupported edition details (e.g. don't say "limited edition" or "remastered" unless confirmed).
After the description, add a separate "Track List" heading followed by the complete, ordered track list.

2. Track listings
Required for every product.
Keep CD disc divisions and vinyl side divisions (Side A/B etc.) where verified. Do not guess vinyl sides from the running order. If the track order is verified but the sides are not, include the track list and flag the missing side information.
Supplied descriptions and track lists may refer to a different edition, so check them. If tracks cannot be verified, omit them and flag this in your report.

3. SEO fields
Factual SEO title and meta description for each product, following our existing store style (examples provided on acceptance).

4. Cover images
Source images from the official record label's website first, then a list of approved websites to source images will be provided. Add the image URLs to the data sheet.
Genuine artwork for the correct release only, ideally at least 600x600 px.
If only smaller genuine artwork exists, use the best verified original and place it on a 600x600 canvas with minimal neutral padding. Do not stretch or crop. Record the original dimensions and note the resize in your report. Do not describe enlarged images as native high-resolution.
Do not reject a good genuine cover just because it is under 1,000 px.
No AI-generated or redrawn covers. No watermarked, unrelated or wrong-edition images.
Front cover first. Add verified back covers, sleeves, labels or disc photos if readily available, but these must not hold up the batch.
Add saved URL to .csv
If no genuine image is available, flag it and continue. Never invent artwork.

Approved sources
Only UK, Dutch (NL) or German (DE) websites may be used (for example .co.uk, .nl and .de domains).
The only exception is the official record label's own website.
Do not use information or images from sources outside these regions.

Tools and pace
We have been doing this work in-house with ChatGPT. About 90% of the data was found successfully, but it is slow: roughly 20-30 minutes per 200 lines.
The successful applicant will need access to multiple AI tools and a workflow that can handle volume. Whatever tools you use, all information must be verified against the correct edition, and unverifiable details must be flagged rather than guessed.
Timeline: the full 60,000 lines must be completed within 5 working days of acceptance. Please confirm in your bid that you can meet this, and explain how (team size, tools, batching).

Deliverables
The completed data sheet in the same structure as supplied, with the new columns filled in.
A report listing: products with omitted or unverified tracks, missing vinyl side information, products with no cover found, and any resized images (with original dimensions).