MerciYanis rewrites its whole platform, about 250,000 lines across 10 repositories, in one day with one engineer and around 30 parallel agents
MerciYanis's CTO describes a one-day rewrite of the company's entire codebase to improve its design and architecture without changing what the platform does. Before starting, the team pushed unit and end-to-end test coverage to 100% on every repository, testing only public contracts so the tests would survive the rewrite. Test files were locked: each iteration ran the unit tests and linter and used git diff to confirm no test had been changed. An internal handbook and the opinionated Perseid framework set the rules agents followed, and when output was wrong the team fixed the prompts and handbook rather than editing code by hand. The post credits Opus 5.5 and GPT 5.6 Sol models but doesn't name a coding tool.
The numbers (as reported)
| Scope | Roughly 250,000 lines across 10 repositories |
|---|---|
| Time and people | One engineer's full-time day, about 30 agents in parallel |
| Tokens | About 1.6 billion |
| Bugs fixed along the way | About 50 bugs and edge cases |
| Regressions | Two small regressions found in QA, both fixed within minutes |
Facts
- Company / subject
- MerciYanis
- Tools
- Unstated
- Codebase
- brownfield
- Work type
- Refactor / maintenance
- Stack
- —
- Published
- 2026-10-26
- Last checked
- 2026-10-28
What the source claims Practitioner (own blog/paper)
Claimed by the builder, writing about their own work. Not checked independently.
- Roughly 250,000 lines across 10 repositories rewritten in a single day by one engineer and about 30 agents 250k lines in 1 day
“All of it: roughly 250,000 lines of code, spread across 10 repositories, in a single day.”
- About 50 bugs and edge cases fixed along the way; two small regressions found in QA and fixed within minutes ~50 bugs fixed; 2 regressions
Caveats
Written by the company's own CTO. The one day excludes the earlier work raising test coverage from an average of about 60% to 100%, which the post says took most of the time. Results are reported from the QA environment; no production figures. Coding tool not named. Stack not stated; the post's code sample looks like TypeScript and its diagram names PostgreSQL and Kafka, so check before filling it in.
More stories by tool: Unstated
Related topics: Refactoring and maintenanceBrownfield codebases