Migrations: 4 sourced coding agent stories
Moving existing code to a new library, platform or architecture, often as many pull requests. Of the 4 stories, 2 were published by a tool vendor and 2 by the people who did the work. 4 involve an existing codebase and 0 started from scratch. Every number links to where it was published.
- Practitioner (own blog/paper) Migration brownfield
Spotify's background coding agent merges 1,500+ migration PRs with 60–90% time savings
Spotify · 2025-11
- 1,500+ Agent PRs merged to production
- 60–90% vs writing by hand Time saved on these migrations
- Hundreds of developers Users
Claude Code, Claude Agent SDK, Honk (Spotify)
- Practitioner (own blog/paper) Migration brownfield
Google estimates ~89% time saved on Joda-Time → java.time migration clusters
Google · 2025-01
- ~89% Time saved on small clusters (estimate)
Google internal migration toolkit (fine-tuned Gemini)
- Vendor-published Migration brownfield
Nubank uses fine-tuned Devin to split a 6M-line ETL monolith, claiming 8–12x engineering-hour efficiency
Nubank · 2024-12
- 6M+ lines, ~100,000 data-class implementations Codebase
- 18 months, 1,000+ engineers Original estimate
- 8–12x Engineering-hour efficiency
Devin
- Vendor-published Migration brownfield
Block uses Claude Fable to orchestrate code migrations of thousands of pull requests, with two humans approving every deploy
Block
- "hundreds or thousands of individual pull requests across multiple repos and millions of lines of code" Migration size
- "we might merge a thousand pull requests orchestrated by Fable" PRs per migration
- "up to dozens of smaller, more cost-effective models like Opus or Sonnet" Worker models
Buzz (Block)