İçeriğe geç
YENİPazar yeri ve trig8n Agent: geliştirdiğinizi satın; planlar ve değişiklikler siz görmeden önce test edilsin. Yeniliklere bak

Kütüphane / İçerik Yazımı ve Üretimi

İçerik Yazımı ve Üretimi

Gemini ve OpenRouter ile Google Docs taslaklarını marka yönergelerine göre inceleyin

Bir kullanıcı Google Doc URL'si içeren bir n8n formu gönderdiğinde, bir AI ajanı içeriği marka dili ve özellik referanslarına göre kontrol eder, ardından sonuçları e-postayla bildirir.

Ne yapar

# Review content pieces in Google Docs This workflow automates the review of content pieces in Google Docs against specific branding and positioning guidelines. It uses an AI Agent to ensure that the content aligns with your brand's voice and correctly references features and documentation. ### Who's it for Content managers, editors, and brand content reviewers who want to ensure consistency and quality in their team's written content, saving time on manual reviews. ### How it works / What it does The workflow is triggered via an n8n form in which the user specifies the Google Doc URL. The workflow downloads the document and extracts the text in .md format. Next, it loads a Data Table containing the brand kit (e.g., tone of voice, product descriptions). The AI Agent then receives the brand kit content as part of a system prompt. The Google Doc content is passed in a user prompt. The AI Agent processes this information. It uses MCP (Model Context Protocol) tools to dynamically search your documentation and blog to verify feature claims and suggest relevant internal links. Finally, the agent structures its feedback (actionable edits, internal links, CTAs, and tone rewrites) and - appends this feedback directly to the original Google Doc; - sends the HTML-formatted report via email ### How to set up 1. Connect your Google Drive and Google Docs credentials. 2. Connect your OpenRouter account for the LLM node. 3. Enter your e-mail address in the Gmail node to receive a formatted HTML report 3. Configure the Kapa.ai and Gitbook MCP tools with your respective endpoints: - No MCP server? The MCP tools let the agent verify features and links against live documentation. If you don't have a Kapa.ai or GitBook MCP endpoint, the workflow still works with one of these workarounds: - Remove the MCP tools and update the prompt. Delete or disable the two MCP nodes and remove the "MANDATORY SOURCING" instruction to call the docs tool. Otherwise the agent will try to call tools that don't exist. - Put your reference material in the Data Table. Add rows such as "Key features and doc links", "Approved internal links", and "CTA banners and URLs". The agent will then pull features and links from the brand kit instead of searching live. - Swap in another tool. Replace the MCP nodes with an HTTP Request Tool pointing at your docs search API or sitemap, or with a vector store tool co

Adımlar

  1. Run Brand Check
  2. When clicking ‘Execute workflow’
  3. Download outline
  4. Extract doc text
  5. Load Brand Kit
  6. Format Brand Kit as one file block
  7. Prepare Brand Kit data for AI Agent
  8. Merge doc text + Brand Kit
  9. Split Out
  10. Populate Data Table with test data
  11. Prepare AI feedback text
  12. Write AI feedback to Google Doc
  13. Convert Report to HTML
  14. Generate Report File

Toplam 22 adım.

Benzer otomasyonlar