phoenix
phoenixPhoenix - Your AI Assistant
Inactive session will be ended in 24 hours
End Session

Backend AI Widget Integration (WordPress + n8n + Claude API)

  • Posted:
  • Proposals: 59
  • Remote
  • #4513555
  • Open for Proposals
Óscar P.GBENGA E.Aadil S.Rohit K.Joshua A. + 54 others have already sent a proposal.
  • 4

Description

Experience Level: Intermediate
Please read before applying

**Timeline:** 10–15 days, or please state how many days you would need.

## Project Overview

I am updating a website with auto data ingestion and normalization platform on WordPress. The system pulls structured data from multiple external sources daily, rewrites content using an NLP API, deduplicates against the existing database, and publishes normalized entries as custom post types. We also need a dynamic content assessment widget that calls the same NLP API to provide structured analysis reports.

## Skills Required

- WordPress development (PHP, custom REST API endpoints)
- n8n workflow automation
- Claude API integration (prompt engineering, structured JSON)
- Docker setup and configuration
- JavaScript (modal widget, localStorage, API calls)
- Google Sheets API
- Apify API
- MailerLite API

## Key Tasks

- Self-host n8n via Docker on staging VPS
- Build 4 custom WordPress REST API endpoints (create, check-duplicate, expire, save)
- Build 4 n8n workflows (daily content ingestion from edX, Udemy, Adzuna, Apify, etc)
- Build AI assessment widget with modal, loading animation, and pre-check logic
- Build WordPress admin settings page (API keys, ad codes, redirect manager, pipeline status)
- Build trackable redirect system (`/go/[slug]`)
- Build save-for-later + reminder system
(Please start your post with...I can execute the WordPress + n8n + Claude API Tasks)

## What We Provide

- Staging WordPress admin login
- Staging VPS SSH access
- Apify sub-user account
- Claude API key ($30 spending limit)
- Google Sheet ID and service account JSON
- MailerLite API key
- All API credentials via n8n Credential Manager
- Briefing Google Doc with all workflow specs and Claude prompt templates

## Evaluation Criteria

- API endpoints function correctly in Postman
- Workflows trigger successfully and create posts without duplicates
- AI widget opens with loading animation and returns correct results
- Redirects track clicks correctly
- Settings page saves encrypted keys

Please include your experience with:
1. WordPress REST API development
2. n8n workflow automation
3. Claude API integration
4. JavaScript modal widgets

State which aspects you can execute well and what it would cost to do the aspects you are very good at.

New Proposal

Create an account now and send a proposal now to get this project.

Sign up

Clarification Board Ask a Question

  • VCONN P.

    Hi Timothy,

    I've a few questions:

    1. Is the AI assessment widget expected to use the same prompt templates as the ingestion workflows, or will it have its own prompt structure?

    2. Approximately how many records are expected to be processed daily across all data sources?

    3. Will the n8n workflows and Docker environment already be provisioned, or are you looking for complete setup and implementation as part of this project?

    4. Is this functionality intended to be delivered as a single custom WordPress plugin or split into multiple modules?

    5. Could you also share your anticipated budget for this project?

    Looking forward to your reply!

  • CYTNHIA W.

    1: As your data sources expand over time, would you prefer the ingestion platform to be configurable by non-technical administrators (allowing new sources and transformations without code changes), or do you expect future sources to be implemented through development? That decision will significantly influence how I architect the workflow engine.

    2: For Claude's structured JSON responses, should the workflows fail immediately if the schema validation doesn't match the expected output, or would you prefer an automatic retry and correction pipeline that re-prompts Claude before the workflow continues? In production AI pipelines, this choice has a major impact on reliability and operational maintenance.