Генерация черновиков ИИ с проверкой качеством и утверждением, затем публикация в Google Docs

## Generate content with AI agents through a QA gate and human approval before publishing Who it is for. Marketing teams that want drafts produced on a schedule but never published without a human and a quality gate. What it does - Webhook or weekly schedule with a list of topics. - Per topic: research (search API), outline, draft, then a QA gate in code: minimum length, required sections, a sources section with two or more links, banned marketing claims. - Passing drafts go to human approval and are published as drafts to Google Docs or the CMS; failing drafts go back with the reasons, never published. - Token usage and cost are accounted per run. Setup 1. Set base in Config or swap the HTTP Request nodes for your search API, Slack and Google Docs or CMS nodes. 2. Set the AI credential on the research, outline and draft steps. Edit the banned phrases and minimum length in the QA gate Code node. 3. Activate. All external systems are reached through HTTP Request nodes against base in the Config node, so the whole flow imports and runs against any REST backend; swap any of them for the native n8n node of your system. Built and tested on sample data by Finamatik: https://finamatik.com/work/content-pipeline-qa-gate Requirements. n8n 1.100 or later; for the AI steps, a model endpoint at base/llm/complete or the OpenAI or Anthropic node in its place; your own systems reachable over HTTPS. How to customise. Every rule lives in one Code or Config node and is named for what it decides; change the thresholds there, not in the branches.