Service
Code Insights
Legacy code intelligence
We analyse legacy, inherited and undocumented codebases and explain what they actually do. Start with a free analysis of up to 500 lines of your code.
The problem
You cannot safely change what you do not understand.
The systems businesses depend on most are often the least understood: undocumented, written in languages few people still use, and maintained by people who have moved on. Changes become risky, migration plans become guesswork, and critical knowledge sits with a handful of people.
What it does
Clarity about a codebase, backed by evidence.
Free initial analysis
Up to 500 lines analysed at no cost and under NDA, so you can judge the value on your own code before committing.
Architecture recovery
How the system is structured, where its components sit and how they interact.
Business logic extraction
The business rules embedded in the code, surfaced and explained, along with what depends on them.
Dependency and data-flow mapping
Modules, jobs, services, databases and external systems, and where a change will ripple.
Risk and quality insight
Fragility, complexity, duplication, dead code and hardcoded assumptions, flagged before they cause trouble.
Plain-English documentation
Findings written for developers, architects and business stakeholders alike.
How it works
- Choose the route: we deploy a guardrailed analysis engine, trained for the language in question, into a pre-production or test environment, or you share a copy of the codebase under NDA, by secure transfer or on encrypted media by courier.
- The engine maps the codebase: its languages, structure, patterns and dependencies.
- It traces business rules, data flows and risks, and our engineers review and verify every finding.
- You receive a detailed report: how the system works, where the risks are, and what should happen next.
What the report covers
Every report is scoped to what you need. A full analysis includes:
- Function descriptions
- Business rules, traced to source
- Data lineage and usage
- Interfaces and integrations
- Batch schedule and job dependencies
- Coding standards and compile-readiness
- Dead code and dead ends
- Loops and control flow
- Complexity and duplication
- Error-handling gaps
- Hardcoded values
- Performance hotspots
- Security vulnerabilities
- Personal data map
- Concurrency and locking
- Date, rounding and precision
- End-of-life platforms and components
- Third-party licences
- Logging and audit trail
- Impact analysis
- Migration readiness by module
- Effort estimate
- Test gaps and suggested test cases
- Documentation drift
01Executive summary
The platform is stable and its settlement logic is sound, but knowledge of it is concentrated in very few places. It can be modernised safely, provided three issues are dealt with first.
- Two programs hold 61% of the business rules and change most often. Any migration should start with a full specification of these two.
- 147 business rules extracted. 23 of them exist only in the code, with no matching documentation.
- 11% of the code is unreachable and can be retired before migration, reducing the effort.
- Two nightly batch jobs write to the same file without locking. This is the most likely cause of the month-end reconciliation errors reported by Finance.
02Dependency map (excerpt)
03Business rules (excerpt of 147)
SETTLE-VAL · line 1182documentedRESERVE-CALC · line 407code onlyPAY-OUT · line 96code onlyRATE-CALC · line 2210outdated value04Risk register (top three)
05Security and data handling
Personal data found in 14 fields across 5 data stores. Two of those fields leave the platform in a nightly extract.
DB-CONN · line 58credentialsCLM-SEARCH · line 311injection riskEXTRACT-N09 · line 140personal dataPREM-TOT · line 882rounding06Migration readiness
07Recommended next steps
- Fix the batch file conflict (critical, low effort).
- Remove credentials from the code, add query validation and encrypt the nightly extract.
- Specify the rating engine and settlement rules in full, with the business owners.
- Plan a staged migration with Code Revive, starting with intake and validation.
Scroll the report. Simulated example with fictional data.
Where it delivers value
Code Insights fits wherever a codebase matters more than it is understood:
- Legacy and mainframe systems, in any language
- Undocumented, inherited or proprietary codebases
- Pre-migration assessment and planning
- Technical due diligence before acquisition or replacement
- Audit, compliance and resilience reviews
- Business-critical systems where the expertise is ageing
The aevris approach
Every codebase is quoted individually. A small sample might produce a concise explanation and risk summary; a larger system can become a knowledge base, an architecture map, a migration plan or an onboarding pack.
Code Insights is usually the first step before Code Revive: it establishes what can move, what should be rewritten and where the real effort sits.
Talk to us about Code Insights
Send us up to 500 lines you are unsure about. We will show you what the first analysis reveals.
