DAG Execution
Definition
DAG execution is a workflow execution mode where steps run concurrently based on their dependency relationships rather than sequentially. Steps that don't depend on each other run in parallel, while dependent steps wait for their inputs. This dramatically reduces end-to-end execution time for complex workflows with independent branches.
How It Works
Each workflow step declares its dependencies — which previous steps it needs outputs from. The DAG executor builds a dependency graph, identifies independent "waves" of steps that can run in parallel, and executes each wave concurrently. If a step fails, its dependents are automatically skipped.
When to Use DAG vs. Sequential
Use DAG mode when your workflow has independent branches that can run in parallel — for example, analyzing a prospect's LinkedIn profile, company financials, and news mentions simultaneously before combining results. Use sequential mode for simple linear chains where each step depends on the previous one.
Related Terms
AI Recipes
Learn what AI recipes are and how they work in JieGou. Recipes are reusable, single-operation AI building blocks with structured inputs and outputs.
AI Workflows
Learn what AI workflows are and how they automate multi-step processes. Workflows chain recipes with branching, loops, approval gates, and parallel execution.