Use cases / Legacy Codebase Migration
Use case
Uzera maps the codebase you are moving off so the team and the agent understand it first, holds the agent to the target patterns as it writes, and re-checks every migration pull request in your CI. A human approves each merge.
The migration, governed end to end
Knowledge Base: understand it first
The target patterns, written as rules. The agent is held to them as it writes, and a violation is denied before anything is saved.
Every push re-checked in your CI
The trend as you go
What Uzera does in a migration
In the order a migration needs them: understand it, set the target, check every step, prove the trend.
Knowledge Base
Uzera indexes the legacy repository where it already lives and explains it: an architecture map, a generated wiki, and plain-language answers about how it works. Across 93 languages.
Guardrails
Write the target patterns as rules in plain English, or adopt a pack. Uzera compiles each rule to an enforceable matcher and rejects it if it can't. The agent is held to them as it writes.
Code Review
Every push to a migration pull request is re-checked in the CI you already run: GitHub Actions, Azure Pipelines, Bitbucket Pipelines, Azure DevOps Server. Findings appear on the pull request.
Code Health
Deterministic scores per scan, no model in the loop: density and trend per repository as the migration moves. Whether the codebase is getting better or worse is a number, not an opinion.
How a governed migration runs
Three steps, repeated for every module you move. Nothing changes in how the team works.
Index the legacy repository. Uzera builds the architecture map and the wiki, and the team and the agent can ask the codebase how it works before anyone changes it. The code itself stays on the machine.
Write the target patterns as rules: the new data layer, the new framework conventions, what is no longer allowed. The agent is held to them while it writes, so the old patterns stop coming back.
Every migration pull request is re-checked in your CI. Findings close themselves when the code changes, a human approves the merge, and Code Health shows the trend after every scan.
Where it fits
Before the budget conversation
No. Your team writes the code. Uzera maps what you are moving off so nobody starts blind, holds the agent to the target patterns, and re-checks every migration pull request.
Modules, dependencies and symbols, read from the code itself rather than from documentation nobody wrote. That becomes an architecture map and a generated wiki you can ask questions of.
Code Health scores every scan per repository, so you can watch the trend move between the start of the migration and the end.
Governance for your coding agents
AI writes your code. You stay in charge.