LinkedIn Content Engine
A content system I built that turns campaign goals, audience notes, and live trend data into a LinkedIn content brief: 10+ posts per run, each matched to a format and a goal, exported as Word documents in under two minutes.
- per content brief
- 10+ posts
- full brief generated
- <2 min
- every post mapped to the objective
- Goal-driven
Node.js, Express, Anthropic SDK, docx, TypeScript
LinkedIn content marketing tends to fail at scale in one of two ways: many posts with no strategy behind them, or a strategy with too few posts to stay consistent. I built the LinkedIn Content Engine to give a team both volume and strategy. It takes a marketing goal and produces a full, structured content brief mapped to that goal, ready to hand off to any writer or publish directly after a final human pass.
What I built
A Node.js pipeline with an Express backend and a form-based frontend. A marketer or content strategist fills in the inputs: campaign theme, target audience, marketing objective (awareness, engagement, lead generation, thought leadership), field notes or product context, repurposing flags for any existing content worth reactivating, and the desired post count. The engine then writes the brief.
- Configurable output volume
- Post count is fully configurable per run. A light week might call for three posts. A product launch or campaign sprint might need ten or more. The generation pipeline scales to match the brief and keeps format variety and strategic coherence across the batch.
- Trend integration
- Before generation runs, the engine pulls the week’s relevant headlines from Google News via RSS, filtered by topic and cached for two hours. The trend data gets woven into the generation prompt so posts connect to something happening in the market that week.
- Objective-mapped generation
- The system prompt encodes a full content strategy layer: hook formats, storytelling structures, engagement mechanics, call-to-action ratios, and tone calibration per audience type. Every post in the brief is mapped to the campaign objective set at input. A thought leadership brief generates differently from a lead generation brief, even with the same source material.
- Dual-document export
- Every run produces two Word documents. The main brief contains all posts formatted and ready to edit or approve. The second is structured as a NotebookLM source document, so the content feeds directly into a research and ideation workflow without copy-pasting. Both are zipped and returned in a single download.
- Usage logging
- Every run logs token consumption, generation time, and campaign metadata behind an API route, so I can track output volume and cost over time.
Why a pipeline rather than a prompt
A raw prompt produces one post, while a pipeline produces a content program, because the pipeline encodes the campaign objective, the audience profile, the format mix, the trend layer, and the structural consistency across every post in the brief. A one-shot prompt loses all of that, which is why the engine pairs a well-built system prompt with a structured generation pass. The LinkedIn Content Engine is that system. A team can point it at a new campaign and get a structured brief they can edit.
What it delivers
A content brief that would take a marketing team an hour to plan, write, and format now takes under two minutes to generate and another ten to review and personalize. From the first draft, the posts come in a mix of formats and follow the week’s news and the campaign objective. Judgment, voice, and final edits stay with the human reviewer.
