← back to examples
ai intermediate

Runic Adaptive Researcher

Dynamic two-phase Runic workflow that hot-swaps between full and slim writing DAGs based on deterministic research richness.

primary showcase ai l2 ai-tool-use runic adaptive

Related guides and notebooks

Tutorial
AI agent with tools
Build tool-aware workflows and orchestration loops.
Livebook
AI agent with tools
Run the companion notebook in Livebook
Source
adaptive demo source
Two-phase adaptive runner script.
Source
adaptive researcher module
Dynamic phase selection and runic.set_workflow usage.
Explanation

What this example teaches

Interactive Demo

Run the agent and inspect state

Source Code

Read the production implementation

What you’ll learn

  • How to split orchestration into phase 1 research and phase 2 writing
  • How deterministic research richness can decide between full and slim phase-2 workflows
  • How runic.set_workflow can hot-swap the DAG while keeping the example fully local and repeatable

Branch behavior

  • full: BuildOutline -> DraftArticle -> EditAndAssemble
  • slim: DraftArticle -> EditAndAssemble

Selection is based on deterministic research summary length in the local phase-1 workflow.

Demo note

This page now runs a real local Runic workflow. The research summaries and article outputs are deterministic fixtures, but the phase selection, workflow swap, node execution, and final artifacts all come from the real strategy pipeline.