Beddel: YAML AI Workflow Designer & Executor

workflow_description: Create a content marketing pipeline that researches a topic, generates blog posts, creates social media summaries, and schedules publication | performance_target: high-throughput batch processing

# Content Marketing AI Workflow ```yaml workflow: name: content_marketing_pipeline version: 1.2.0 description: Automated content creation from research to publication config: timeout: 1800 # 30 minutes retry_policy: max_attempts: 3 backoff_factor: 2 error_handling: continue_on_failure steps: - name: topic_research type: ai_task model: gpt-4 prompt: "Research comprehensive information about {topic}. Include statistics, trends, and expert insights." inputs: topic: ${workflow.input.topic} outputs: research_data: research.json timeout: 300 - name: content_generation type: ai_task model: claude-opus prompt: "Write a 1500-word blog post about {topic} using this research: {research_data}" inputs: topic: ${workflow.input.topic} research_data: ${steps.topic_research.outputs.research_data} outputs: blog_post: content.md depends_on: [topic_research] - name: social_media_creation type: parallel steps: - name: twitter_thread type: ai_task model: gpt-3.5-turbo prompt: "Create a 5-tweet thread summarizing: {blog_post}" inputs: blog_post: ${steps.content_generation.outputs.blog_post} outputs: tweets: twitter.json - name: linkedin_post type: ai_task model: gpt-3.5-turbo prompt: "Create a professional LinkedIn post from: {blog_post}" inputs: blog_post: ${steps.content_generation.outputs.blog_post} outputs: linkedin_content: linkedin.md depends_on: [content_generation] ``` ## Execution Instructions 1. Install: `pip install beddel-workflow` 2. Run: `beddel execute content_marketing_pipeline.yml --input topic="AI in Healthcare"` 3. Monitor: `beddel status workflow_id` ## Scaling Optimizations - Parallel social media generation reduces total time by 40% - Retry logic handles API rate limits automatically - JSON output format enables easy integration with publishing platforms