Building My LinkedIn Automation Web Application (From OAuth to Live Posting)

Source: Dev.to

Overview

After spending time understanding OAuth and social media APIs, I decided to build a real LinkedIn automation system—a web application where a user can:

1. Connect their LinkedIn account
2. Write a post (text or image)
3. Click Publish
4. Have the post go live automatically

The idea is simple, but the execution is complex.

OAuth 2.0 Implementation

Steps I Implemented

• Created a LinkedIn Developer App
• Configured the Redirect URI
• Generated an Authorization Code
• Exchanged the code for an Access Token

Receiving the first access token was a pivotal moment—it confirmed that the authentication pipeline was working.

Simplified Flow

User → LinkedIn Authorization Page → Grant Access → Redirect with Authorization Code
      → Server exchanges Code for Access Token → Token stored for API calls

Testing the UGC Post API

With the token in hand, I tested the LinkedIn UGC Post API using hard‑coded text. After refreshing my LinkedIn profile, I saw a real post created from the application. This proved that the integration was functional beyond a simulation.

Dynamic Posting

To move beyond hard‑coded content:

• Built an HTML form
• Connected the frontend to a Flask backend
• Allowed users to enter custom text

Now the application is interactive, letting users write their own posts.

Image Upload Workflow

Posting images to LinkedIn requires a multi‑step process:

1. Register media upload → receive an upload URL
2. Upload the image binary to the URL
3. Attach the media asset to the post

Careful sequencing and correct JSON payloads were essential. After implementation, the app supported:

• ✅ Text posts
• ✅ Image posts

Frontend UI

Initial State

• Plain HTML

Upgraded with Bootstrap

• Modern login page
• Clean post‑creation form
• Responsive layout
• Success confirmation page

The UI transformation turned the project from a “backend experiment” into a polished web application.

Challenges & Errors

• Redirect URI mismatch errors
• Unauthorized scope issues
• 403 Access Denied errors (LinkedIn blocking /v2/me endpoint)
• Decoding id_token
• Handling 500 Internal Server Error

Each issue forced deeper debugging and a better understanding of LinkedIn’s API behavior.

Project Evolution

1. OAuth testing
2. Static text posting
3. Dynamic user input
4. Image upload integration
5. UI enhancement
6. Error‑handling improvements

Each phase added complexity and deeper knowledge.

Skills Gained

• OAuth 2.0 Authorization Code Flow
• Access token exchange & id_token decoding
• REST API integration (including media upload)
• Backend development with Flask
• Frontend‑backend communication
• Real‑world API error debugging
• Secure integration with third‑party platforms

Future Plans

• Multi‑user system with database storage
• Per‑user token management
• Post scheduling
• Multi‑platform integration (e.g., Twitter, Facebook)
• Deployment to a cloud environment
• Analytics dashboard

The project has strong potential to evolve into a full automation platform.