Guides
Guides
Evergreen reference pages that are updated as the product changes, unlike dated blog posts. Start with how to choose a code-intelligence MCP server, then read the guide for your language.
Choosing a server
Language guides
What Maguyva extracts for each language, and the agent workflows that help most in that kind of repository.
-
Python
Useful when Python codebases sprawl across services, scripts, type stubs, and operational glue.
-
TypeScript
Relevant when your repo mixes app code, libraries, API clients, tests, and config across multiple packages.
-
JavaScript
Useful when a repo mixes CommonJS, ESM, task scripts, tests, and older application code.
-
Go
A good fit for service-heavy repos where handlers, packages, and operational code need to stay easy to trace.
-
Rust
Relevant when you want AI assistance in Rust without losing the structural guardrails that make Rust worth using.
-
Java
Useful when the repo is full of Spring services, internal frameworks, and code that has survived several organizational eras.
-
C#
A good fit when .NET repos have grown large enough that safe AI assistance depends on real structure.
-
PHP
Relevant when a PHP codebase is still shipping revenue but nobody wants an agent freelancing through it.
-
Terraform
Useful when infrastructure code is part of the repo context and not something the agent should hand-wave past.
-
COBOL
This is where broad language support stops being a stat and starts becoming a modernization tool.
-
Fortran
For numerical code that outlived its authors, a call graph is the first piece of documentation.
-
ABAP
Before an S/4HANA move, you need to know what your custom code actually depends on.
-
Ada
Ada was built for systems that must not break. Changing them deserves the same care.
-
Pascal
Delphi and Free Pascal applications run for decades. Understanding them should not take that long.
-
Perl
The Perl that still runs your billing, reporting, or deploy scripts deserves better than grep.
-
Kotlin
Kotlin hides a lot of wiring in receivers and lambdas. The graph puts it back.
-
Swift
In Swift, behaviour lives in protocols and extensions. An agent needs to see both.
-
Ruby
Ruby reads like prose. The dependencies behind it still need a map.
-
C
In C, the call graph is the architecture. An agent should read it before it edits.
-
C++
C++ hides the real call target behind namespaces, overloads, and virtual dispatch. The graph helps an agent find it.
-
Scala
Scala packs a lot of meaning into little syntax. The graph unpacks it.
-
Dart
Flutter apps grow wide fast. The graph keeps the wiring visible.
-
Lua
Lua builds objects out of tables. The graph shows which tables became classes.
-
R
Analysis code outgrows its first script. The graph shows what depends on what.
-
PowerShell
Automation estates sprawl across scripts. The graph shows which function runs what.
-
Groovy
Build logic is code too. The graph treats it that way.
-
Elixir
Elixir's structure lives in macros. The graph reads them as structure.
-
Haskell
Haskell's types say a lot. The graph shows who relies on them.