eBay Listing Automation Developer

I'm seeking a skilled developer to create a listing automation tool for eBay. The primary focus is on automating product uploads.

Key Requirements:
- Develop a tool to automate product uploads to eBay.
- Ensure the tool is user-friendly and integrates seamlessly with eBay's existing platform.

Ideal Skills and Experience:
- Proven experience in developing automation tools, especially for e-commerce platforms.
- Strong understanding of eBay's API and listing requirements.
- Ability to create intuitive user interfaces.
- Background in inventory management systems is a plus.
Developer Brief — "The Reseller's Studio" → eBay-linked, high-volume listing platform

1. Summary

I run a vintage clothing resale business listing on eBay UK at volume — around 100 listings a day. I have a working tool, The Reseller's Studio, that prepares an eBay listing from photos of a garment (title, description, item specifics, measurements, SKU, price). It currently runs as a web app and produces everything ready to copy and paste into eBay.

I want a developer to build this into a proper platform that:


links to my eBay account and publishes listings live (not copy/paste, and not draft-only — I want them to go live, ideally with an option to hold as a draft when I choose);
automates the product image step via an image-AI API (ghost mannequin / background removal) so it happens in-flow rather than via a separate manual tool;
suits high volume and a small team — I want to delegate the actual deployment to an assistant.


Given the volume and team use, I want this as a web application used in a desktop browser, responsive so it also works on a phone. See §6.

The existing front end and its logic can be reused; the main new work is the eBay integration, the image-API integration, a backend, multi-user/queue handling, and bulk/live publishing.


2. What exists today (the starting point)

A self-contained web app — index.html plus manifest.webmanifest, a service worker and icons. Plain HTML/CSS/vanilla JavaScript, no build step, no framework. (I will provide the full source as a zip.) Four steps:


Step 1 — Garment details: two-part SKU (Box + Item, joined into one custom label), price, department, category, era, condition, material (multi-select; several = a blend), size and size type, measurement diagrams with labelled arrows per garment type (plus a user-uploadable custom diagram), and seller notes.
Step 2 — Photos: gallery (add, reorder, light auto-enhance, up to 24), export/share of 1600×1600 JPGs.
Step 3 — Ghost mannequin: currently a manual hand-off to Snappyit. To be replaced — see §5.
Step 4 — AI listing: Anthropic API generates title (≤80 chars), item specifics and description from the photos, reading the brand/care label for brand, size and fibre composition. Includes a "save to batch / export CSV" worklist (SKU-first columns) as an interim measure.



3. What I want built

Core


eBay account linking — authorise the platform to act on my eBay account (one-time consent, persists).
Publish live to eBay — create a live listing populated with title, description, item specifics/aspects, price, condition, category, photos, and the SKU as eBay's "Custom label (SKU)". Provide a per-item option to save as a draft instead of going live.
Field mapping — app fields → eBay fields: category, condition values, item specifics → aspects, two-part SKU → custom label.
Photo upload — upload the gallery images as listing pictures (1600×1600, in order).
Sold-order SKU lookup — retrieve sold orders and show the item's SKU/custom label so I can find the right physical box.


Image processing (replaces the manual Snappyit step)


Integrate an image-AI API to produce the ghost-mannequin / clean product image automatically as part of the flow. Default candidate: Photoroom's Ghost Mannequin API (part of its Image Editing API). Please keep the integration provider-agnostic so it can be swapped, and before committing, trial the chosen provider on a sample of my actual garments (which include knitwear, silk and dark/complex pieces, where general-purpose tools can be weaker) against at least one apparel-specialist alternative (e.g. a fashion-focused ghost-mannequin API, or fal.ai-hosted models). Account for per-image credit costs at ~3,000 images/month.


High volume & team use


Multiple users — at least two roles: me/staff preparing items, and an assistant publishing them. Simple per-user login.
A stored, shared queue of prepared listings (saved centrally, not just one browser), shown as a list/grid, assignable, reviewable and publishable by the responsible person. The queue should lead with the SKU as the first/identifying column, then title, price, status.
Built for ~100 listings/day — fast keyboard entry, batch actions (publish selected), and a publishing route suited to volume (see §4).


Security


Move the Anthropic key, the image-API key and the eBay credentials/tokens to a backend; the client must never hold secrets. The backend handles eBay OAuth, token storage/refresh, and all third-party calls.



4. Technical notes

eBay (developer will know current details):


eBay developer account + application keyset (App ID / Cert ID / Dev ID); sandbox then production. Production listing access needs eBay approval — factor in lead time.
OAuth 2.0 (authorisation-code grant) user tokens with refresh, scoped for selling.
Publishing — please advise which suits ~100/day live:

Sell APIs (per item): Inventory API (inventory item keyed by SKU → offer → publish), Account API for business policies, Fulfilment API for sold-order SKUs, image upload to eBay Picture Services.
Bulk via File Exchange / Seller Hub CSV (Custom Label/SKU first column) for many at once; needs hosted image URLs and correct category IDs.



Marketplace eBay UK (EBAY_GB), currency GBP.


Image API: default Photoroom (Ghost Mannequin / Background Removal endpoints; paid plan + credits). Keep provider-agnostic; see §3.6.

Backend: required (e.g. Node or Python) — eBay and the image API are server-to-server and secrets must not be in the client.


5. Image step — replacing Snappyit

The current Step 3 hands off to Snappyit, which has no public API and so cannot be automated. Replace it with an API-based image provider (default: Photoroom Ghost Mannequin API) integrated into the flow, per §3.6. Snappyit should be dropped from the build.


6. Platform / architecture


A web application in a desktop browser for day-to-day work, responsive for phone use. (The current phone-style build was for solo on-the-go use; the team/volume workflow is better as a web app.)
Hosting for front end and backend (a small managed host is fine). Include setup and a brief written handover.
A Play Store native wrapper is not required; quote separately only if recommended.



7. What I will provide


Full source of the existing app (zip).
My eBay account; I will register/authorise the eBay developer account and set up business policies (with guidance).
Anthropic API billing, and the image-API account/billing once a provider is chosen.
A domain (optional) and a small monthly hosting budget.
Decisions on defaults: fixed-price vs auction, default postage and returns, live vs draft per item, and user roles/logins.



8. Out of scope (unless separately quoted)


Cross-listing to Etsy, Vinted, etc. (possible future phase).
Native Play Store app (optional).



9. Note on the existing interim build

To bridge the gap, the current app already has a "save to batch / export CSV" worklist (SKU-first columns: SKU, title, price, condition, category, material, size, measurements, item specifics, description; photos shared separately). Treat this CSV shape as the intended data model for the shared queue and any bulk upload.


10. Deliverables & acceptance

Deliverables: working web app (front end + backend) deployed; eBay app keyset configured; image-API integrated; multi-user login and a shared listing queue; live publish-to-eBay (per-item and/or bulk) with a draft option; source code; short written handover (run, deploy, renew credentials).

Acceptance test: I prepare an item (photos, details, SKU, price); the chosen image API produces the ghost-mannequin/clean image automatically; the item is saved to the shared queue; my assistant opens the queue, reviews it, and publishes it; a correctly populated live eBay listing is created — right title, photos in order, price, condition, item specifics, and the SKU as the custom label — and that SKU is visible to me when the item sells.