How it works overview
Phase 03

Execute & Learn

Run once. Route smarter next time.

Overview

Execution closes the loop: Devtor runs the approved plan, streams progress in the terminal, and compares actual token usage and cost against the pre-flight estimate.

In depth

After each task, patterns are cached — task shape, files touched, model used, tokens consumed. Similar future prompts route faster and cheaper. Cross-user learning (when enabled) lets resolved failures become organizational knowledge instead of repeated mistakes.

What happens

  1. 1Execute against the approved plan and context
  2. 2Stream intermediate steps to the terminal
  3. 3Record actual tokens, latency, and cost
  4. 4Cache routing patterns for similar tasks
  5. 5Surface learnings for team-wide reuse

Terminal output

terminal
✓ Completed in 3.2s
✓ Tokens used: 1,847 / 2,400
✓ Cost: $0.006
✓ Pattern cached
✓ 12 similar tasks will benefit

Benefits

  • Actual vs. estimated cost tracking per run
  • Pattern cache speeds up repeat task types
  • Interactive think mode for mid-run corrections
  • Organizational learning from shared resolutions

Ready to orchestrate?

Installation guide and CLI — coming soon.

Get started