[Hiring] Expert to complete Automated YouTube News Workflow (n8n + OpenAI)

Jobs
,
1

Detailed Project Overview

1. Completed Work:

  • Infrastructure: I have a self-hosted n8n environment running via Docker/Portainer (gbn-automation stack) on a VPS. My instance is live and stable.

  • Workflow Architecture: The core logic is already mapped out in n8n. I have placed the following nodes: Schedule Trigger, RSS Read, Limit, OpenAI (Message a Model), and Google Sheets (Append Row).

  • Integrations: Basic credentials for Google Sheets and OpenAI have been initiated.

2. Assistance Needed:

  • Data Mapping & Connections: I am experiencing some UI/UX difficulties in connecting the RSS Read output to the Limit node and ensuring the data schema flows correctly into the OpenAI node.

  • Prompt Engineering: I need help refining the OpenAI prompt to generate high-quality, YouTube-ready news summaries from the RSS descriptions.

  • Data Formatting: Ensuring the output from the AI is correctly parsed into specific columns in Google Sheets (e.g., Date, Title, Summary, Source URL).

  • Error Handling: Implementing “Wait” or “Error” nodes to ensure the workflow doesn’t break if an RSS feed is temporarily down or the AI API hits a limit.

  • Future Expansion: Guidance on eventually connecting this to a video creation tool or the YouTube API directly once the script-saving process is perfect.

Objective: I am looking for an expert to spend a few hours with me (or asynchronously) to “clean up” the connections, verify the data mapping, and ensure the workflow runs 100% reliably from start to finish.

4 curtidas
3

Hi,

I’ve reviewed your current setup. The difficulties you’re facing with connecting RSS to OpenAI usually stem from how n8n handles item arrays (JSON mapping). I can fix your data schema, refine your prompts for YouTube-ready scripts, and make the workflow 100% reliable.

How I will help you:

  • Data Mapping: I’ll implement Item Lists / Split in Batches nodes to ensure every RSS entry is correctly parsed and sent to OpenAI without data loss.

  • Prompt Engineering: I will design a structured System Prompt for OpenAI to generate high-retention YouTube scripts (Hooks, Key Facts, CTAs) instead of dry summaries.

  • Reliability: I’ll add Error Trigger nodes and “Wait” logic to handle API rate limits and dead RSS feeds gracefully.

  • Future-Proofing: I can advise on connecting this to InVideo / HeyGen API or YouTube API for a fully autonomous video factory.

Terms:

  • Format: I can work asynchronously (you send the JSON, I return it fixed) or via a live session.

  • Estimated Time: 2–4 hours of focused work.

  • Rate: $40

Portfolio & Contact:

2 curtidas
4

Well thats great , can we have a whatsapp call +8801618884435 and to explain you my project requirement..

5

I can help you with my prior experience in AI automations, and I have also built my own saas cvscrenner.com and kick-started the AI blog posts machine to write a blog post every day with seo upbringing format hope this knowledge and exp in this can help you

8

Hey,

You don’t have a “connection issue” — you have a pipeline design problem.

What you’ve built is a working demo, but it’s not structured for clean data flow or consistent AI output. That’s why you’re hitting friction between RSS → OpenAI → Sheets.

I’ve built similar pipelines, and the fix is straightforward once the architecture is corrected:

  • Proper item handling (Split in Batches + normalized JSON structure)

  • Controlled prompt design (so output is predictable and column-ready)

  • Clean parsing layer before Google Sheets (no messy AI dumps)

  • Fail-safe logic for RSS failures + API limits (so it doesn’t silently break)

More importantly — if your goal is YouTube, the current “summary → sheet” flow is incomplete.

You actually want:
RSS → Structured Script (Hook + Segments + CTA) → Storage → Video/Voice Layer

Otherwise you’ll just generate boring summaries that don’t convert.

I can:

  • Fix your current workflow fast

  • Turn your prompts into YouTube-ready scripts (not generic summaries)

  • Structure the output so it’s ready for video automation later

Portfolio (relevant builds + demos):
https://muhammad-ai-automations.notion.site/Muhammad-Bin-Zohaib-AI-Automation-Projects-29da292a241380f889c2e337a134c010

If you send your workflow JSON, I’ll point out exactly where it’s breaking and what needs to change.

– Muhammad

9

Hello @ngncore

I made this video specifically for you: [Hiring] Expert to complete Automated YouTube News Workflow (n8n + OpenAI) - ngncore | Loom

I’ve been building n8n workflows for 6 and 7-figure businesses for 2 years and the video shows exactly a few of them.

Also here is my portfolio: Fran´s Portfolio - Google Präsentationen

Shoot me a message and let’s get started. franmolina.cu@gmail.com

Fran

P.S.: The video is 2 minutes. Worth it.

12

Hey — this is exactly the kind of project we do at Neonotics.

We build and debug n8n + OpenAI workflows for clients. Self-hosted n8n on Docker/VPS is our default environment, so no ramp-up needed.

A few questions to scope this quickly:

- What’s the current workflow doing vs. what it should be doing?

- Where is it breaking — is it a node config issue, API auth, or logic flow?

- What’s your target output format (YouTube video descriptions, scripts, news summaries)?

We can typically complete and test a workflow like this in 2–4 hours once we have access. Happy to do a quick async review of your current workflow first so you know exactly what’s needed before committing.

— Neonotics

13

Hi, This is a strong match for my background. I work with n8n, OpenAI, RSS workflows, and Google Sheets automations, so I can help clean up your existing setup and make it run reliably.

From your overview, the main need is fixing the node connections and data mapping, improving the prompt for better YouTube-ready summaries, formatting the output into clean Google Sheets columns, and adding error handling so the workflow does not break on feed or API issues.

:telephone_receiver: Book a quick call:Calendly - Automaxion

15

Hey! This is exactly what I’ve been building.

I recently completed an automated news channel workflow in n8n — it handles content discovery, summarisation via OpenAI, and publishing. Happy to share a screen recording of how it works.

I’m a sysadmin with hands-on n8n experience (self-hosted), comfortable with webhooks, REST APIs, and multi-step agent flows. I also have a badge for completing the official n8n course.

What’s the current bottleneck in your workflow — the content sourcing, the AI summarisation, or the publishing step? I’d like to understand the specific gap before proposing anything.

16

Hey Mihail, this is right in our wheelhouse — we build and complete n8n + OpenAI workflows regularly, including content automation pipelines. We’d be happy to take a look at what you have so far and get it across the finish line. Feel free to DM us with the details or share what’s currently blocking you.

17

Hi @ngncore,

It sounds like you have a great foundation built already! Getting the self-hosted Docker instance live and the core nodes mapped out is usually the hardest part. I can definitely jump in and help you clean up the data flow to get this running 100% reliably.

Here is how we can tackle your specific roadblocks:

  • Data Mapping & Connections: I will help you properly configure the item loops (likely using the Item Lists node or precise expression syntax) so the RSS Read output passes cleanly through the Limit node and iterates correctly into OpenAI.

  • Prompt Engineering & Data Formatting: To ensure your Google Sheets append perfectly every time, we will refine the OpenAI prompt to use Structured Outputs (JSON mode). This forces the AI to always return exactly what you need (Date, Title, Summary, Source URL) in a predictable format, eliminating parsing errors.

  • Error Handling: I can set up robust error management using the Error Trigger node and Wait nodes to catch API rate limits or RSS downtime, ensuring the workflow pauses and retries rather than crashing.

  • Future Expansion: When you are ready, adding a node to push these summaries directly to a video creation API or the YouTube API is very straightforward once this foundational data structure is solid.

As an AI Automation Specialist and R&D Lead, I architect complex, multi-step n8n automation workflows integrating AI models and external APIs daily. I am also fully comfortable navigating Docker-based environments.

I am available to hop on a quick screen share to do this live, or we can handle it asynchronously, whichever you prefer.

You can check out some of my automation work here:

Let me know if you would like to connect!

18

Hi @ngncore,

YouTube automation with n8n + OpenAI is interesting. I have built similar AI content pipelines on n8n.

What I can bring:

  • Experience building n8n workflows that integrate OpenAI for content processing
  • Webhook orchestration, HTTP nodes, and API integrations
  • Production-ready workflow design with error handling and monitoring
  • Fast turnaround: first working update within 12-24h

Happy to review what you have built so far and propose a path to completion. DM or email: deluca1980fabrizio@gmail.com

Fabrizio — Altiora

20

Your RSS Read output is not connecting to Limit or flowing correctly into OpenAI — is this still open?

If it is and stays written, I’ll reproduce that path with a redacted RSS fixture, correct connections and mapping, then return the export, before/after evidence, and rollback steps. Acceptance: the agreed item reaches Limit and maps the expected fields into the OpenAI input.

Proofs: pilot sample The finalization pilot — one existing workflow, one named defect, 80 USD, 3 days · PERLY — tested internal work n8n workflow examples — code and passing tests · PERLY — my template in the n8n library Triage inbound Gmail with Anthropic Claude, Google Sheets, and Slack alerts | n8n workflow template — my MIT-licensed test runner https://www.npmjs.com/package/n8n-testkit (these are my builds, not client cases).

80 USD fixed, delivered in 3 working days after the milestone is funded. It excludes live credentials, production activation, external-API tests, prompt redesign, and other paths.

Fund the milestone and send the workflow export, one redacted RSS item, current output, and expected OpenAI input.

21

Hi @ngncore,

Is this project still open?

I work with n8n business automations involving Google Sheets, data mapping, API integrations, status tracking, and workflow debugging. I haven’t built this exact YouTube news pipeline before, so I’d prefer to start with a small, clearly scoped paid fix rather than overstate my experience.

For the first milestone, I can focus on:

  • RSS Read → Limit data flow
  • mapping the RSS fields correctly into OpenAI
  • structured OpenAI output for Date / Title / Summary / Source URL
  • mapping the result into Google Sheets
  • basic error handling and end-to-end testing

Price: $60 fixed
Delivery: 2–3 working days after receiving the workflow JSON and sample data.

I prefer to handle the first review asynchronously. You can send me the workflow JSON, one sample RSS item, the current output, and the expected result, and I can first identify where the flow is breaking.

My workflow portfolio:

If the core workflow is fixed successfully, we can discuss prompt improvements or future YouTube/video automation separately.

Best,
Tao Li

22

Hi Perly, this is within my wheelhouse. I work with n8n, Node.js, APIs, RSS workflows, OpenAI integrations, and data mapping.

I can reproduce the issue with a redacted RSS fixture, trace the data through the Limit and OpenAI nodes, fix the connection/mapping, and provide before/after evidence plus rollback steps.

The acceptance criteria you described are clear. If the workflow export and sample RSS item are available, I can review the path and confirm the fix approach before starting.

23

Hi @ngncore - I run the same shape of pipeline every day, so I can be specific rather than general.

My own setup: self-hosted n8n on a Hetzner VPS in Docker, 103 workflows, 74 live 24/7. One of them pulls nine RSS and JSON feeds on a schedule, sends batches to an LLM, and writes structured output into a table. That is your workflow with different sources.

On your four blockers:

RSS into Limit into OpenAI. RSS Read emits one item per article, so Limit is doing what it says, and the mapping problem is usually further down: the OpenAI node needs one field, and it is easy to hand it the whole item instead of contentSnippet or description. Worth deciding first whether you want one model call per article or one call per batch - batching is far cheaper per article and it changes how the node is wired.

Prompt for YouTube-ready summaries. What made the difference for me was forcing strict JSON with named fields instead of prose, at a low temperature. Prose has to be re-parsed downstream and it drifts; JSON either parses or it does not.

Output into Sheets columns. Follows from the above: once the model returns named fields, the Sheets node maps field to column and stops guessing. Add a validation step that drops a bad row rather than writing a half-parsed one.

Error handling. This is the part I would push hardest on, because these failures are silent. I have had a workflow hit a provider daily token cap and stamp a default value on every row while reporting success - green execution, garbage data. So: a real error workflow, a second provider behind the first, and a check that measures the output rather than the exit code. A dead RSS feed should skip that source, not kill the run.

Fixed price agreed before I start, working in your instance or a copy, whichever you prefer.

Based in Spain, remote, English. Happy to screen-share and walk through one of my live workflows first so you can judge the build quality before committing to anything.

24

Hi ngncore — is this project still open? I can handle a tightly scoped first milestone: diagnose the RSS Read → Limit → OpenAI mapping, produce structured Date/Title/Summary/Source URL output for Google Sheets, add basic validation/error handling, and return the corrected workflow plus test notes. I would work asynchronously from a sanitized workflow JSON and one sample RSS item. I do not claim prior client deployment of this exact pipeline, but the scope matches my hands-on n8n troubleshooting and AI-assisted content workflow work. Proposed first milestone: USD 50 fixed, 2–3 working days, starting only after payment protection or a funded milestone is confirmed.

25

Hi @ngncore,

I can help you resolve the pipeline bottlenecks in your n8n workflow and get your YouTube news automation running reliably.

How I will fix and optimize your workflow:

Data Mapping & Schema Fixes: Correct the item array formatting between the RSS Read, Limit, and OpenAI nodes using Item Lists / data transformation nodes so your JSON schema passes cleanly without dropping items.

YouTube-Ready Prompt Engineering: Refine the OpenAI system prompt to produce structured, high-engagement YouTube news scripts (Hook, Main Facts, CTA) rather than basic summaries.

Structured Data Formatting: Parse the AI response into distinct output fields (Date, Title, Script Summary, Source URL) mapped directly to your Google Sheets columns.

Fail-Safe Error Handling: Add Error Trigger and retry/wait nodes to gracefully handle temporary RSS downtime or API rate limits without crashing the execution.

Future Scaling: Set up the output structure so it can easily feed directly into video generation tools (like InVideo/HeyGen) or the YouTube API later.

Execution:

We can complete this asynchronously via JSON exchange or through a quick screen-share session to review the fix together.

Let me know if you’d like to share your workflow JSON so I can take a look!

Best regards,

26

Hi @ngncore,

I can help you fix and finalize your n8n workflow. The array mapping issue from RSS Read to OpenAI/Google Sheets is a common JSON structure mismatch in n8n, which can easily be resolved using proper Item Lists / Code nodes along with error-handling logic.

How I can assist you:

Data Flow & Mapping: Properly structure JSON output from RSS to feed cleanly into OpenAI and Google Sheets columns.

Prompt Engineering: Refine the system/user prompts for high-quality YouTube news scripts.

Error Handling: Implement error triggers, retry conditions, and wait nodes for API resilience.

I am ready to collaborate asynchronously or jump on a call to get this working reliably for you. Looking forward to connecting!

27

Five months and nineteen replies later, my guess is nobody actually finished this for you — so here’s a zero-risk way to close it.

Send me the workflow JSON. I’ll return it fixed and tested the same day: RSS mapped one item per article (the collapse to a single row comes from the OpenAI node returning one item — a Set node after the model fixes it), explicit Date / Title / Summary / Source URL columns so Sheets never guesses, retry-with-backoff on the fetch and model nodes, and an error path that logs a dead feed instead of killing the run.

You look at it running before any money moves. $120 if it works, nothing if it doesn’t.

One question so I fix the right thing: one summary per article, or one combined digest of all of them? Those are two different builds.

Async — you get the JSON back plus a short screen recording of what changed. GitHub: dkautomation23

Dmytro