1.4.x already does a lot — but not every language and framework is understood at the same depth.
This article explains exactly how far ChatCode 1.4.x support goes: what is currently file-level read/edit support, what already has Project Brain and specialized workflows, the remaining limitations, and the priorities being developed for the 1.5.x line.
Three support levels to distinguish.
This is the most accurate way to understand ChatCode's current capabilities.
Read / search / edit
AI can read files, search, retrieve context, and edit project content according to the permissions you grant. This is the broadest support layer.
Project Brain deep understanding
ChatCode analyzes symbols, imports, references, and part of the call graph using semantic parsers. Not every language in 1.4.x has this layer yet.
Specialized workflows
ChatCode detects runtimes/frameworks, installs dependencies safely, runs build/test tasks, or launches previews through controlled profiles.
Languages: 1.4.x is strongest with Rust, Python, TypeScript, and JavaScript.
These currently have deep semantic parsing in Project Brain.
| Language | Read / edit / search | Semantic Brain | Notes |
|---|---|---|---|
| Rust | Excellent | ✅ Deep | Symbols, imports, references, call relationships; strong Cargo tasks. |
| Python | Excellent | ✅ Deep | Good symbol/reference support; runtime + venv are available, while task profiles remain limited. |
| TypeScript / TSX | Excellent | ✅ Deep | Well suited to modern web projects; understands imports/references and package structure. |
| JavaScript / JSX / MJS / CJS | Excellent | ✅ Deep | Node/package workflows are well supported. |
| PHP | Good | ⚠ Not deep yet | Runtime, Composer, and Laravel support are already solid, but PHP AST/semantic depth has not caught up yet. |
| Go, Java, Kotlin, C#, C, C++ | Good | ⚠ Not deep yet | Brain still indexes/searches these files and AI can edit them; dedicated semantic parsers are not available yet. |
| Vue, Svelte | Good | ⚠ Partial | Files are indexed; frameworks can run through Node/Vite, but component semantics are not deep yet. |
| HTML, CSS/SCSS, SQL, GraphQL, XML, Shell, PowerShell, JSON, YAML, TOML, Markdown… | Good | — | Very useful for search/context/configuration, but these are not primary AST languages in Brain. |
Frameworks and runtimes: the web stack already has a solid foundation.
The 1.4.x Environment Doctor can detect multiple project types and prepare environments more safely than letting AI run arbitrary commands.
| Project / Framework | Detection | Dependencies | Automatic preview |
|---|---|---|---|
| Laravel | ✅ | Composer + Node when needed | ✅ Artisan serve |
| PHP generic | ✅ | Composer | ✅ PHP built-in server |
| Next.js | ✅ | npm / pnpm / yarn | ✅ Via dev script |
| Nuxt | ✅ | npm / pnpm / yarn | ✅ Via dev script |
| Vite | ✅ | npm / pnpm / yarn | ✅ |
| Astro | ✅ | npm / pnpm / yarn | ✅ |
| SvelteKit | ✅ | npm / pnpm / yarn | ✅ |
| Node.js generic | ✅ | npm / pnpm / yarn | ✅ If scripts.dev exists |
| Python | ✅ | venv + pip / requirements.txt | ⚠ No broad framework profile yet |
| Rust | ✅ | Cargo | ⚠ Project-dependent |
| Static HTML | ✅ | None required | ✅ |
Laravel has a safe bootstrap
If .env is missing but a template exists, ChatCode can create a local environment, use project-local SQLite, generate APP_KEY, and run migrations. If .env already exists, ChatCode does not automatically overwrite your credentials/database configuration.
App-managed runtimes
1.4.x already has managed runtimes for Node, PHP, Python, and Composer. Rust/MSVC still depends more heavily on the system toolchain.
Build / test: Cargo and package.json are currently the two strongest paths.
ChatCode intentionally uses an allow-listed task model instead of exposing an arbitrary shell to AI.
Rust / Cargo
Check, test, Clippy/lint, build, and bench; with filters, workspace/package selection, features, and release options in the typed task runner.
Node / package.json
Test, lint, typecheck, check, and build through declared scripts; npm, pnpm, and yarn are detected from lockfiles.
Capabilities already present in 1.4.x beyond languages/frameworks.
This is what makes ChatCode different from a simple editor or file-reading plugin.
Project Brain
Index lớn, incremental update, tìm context, symbol search, reference và impact analysis.
Documents
Read and extract PDF, DOCX, XLSX/XLSM, PPTX, and OpenDocument files; original files can be sent to AI when native processing is needed.
Per-project permissions
Read, Write, Full file, Task, and Git-write are separated; sensitive paths and symlink escapes are blocked.
Git & GitHub
Status, diff, branch, stage, and commit; private GitHub workflows include guards to reduce accidental public exposure or pushes.
Remote servers
SFTP/FTPS/FTP through project-bound profiles; SSH execution only when both the profile and local permissions allow it.
Preview & remote work
ChatCode can own the dev server, verify health, and create scoped preview URLs instead of sharing arbitrary localhost ports.
A serious storage issue was identified in 1.4.12.
This is why 1.4.13 is being used for a hotfix instead of waiting for 1.5.
1.4.13 hotfix direction
Reuse the project's standard Cargo cache where appropriate, avoid duplicating a dependency store just because ChatCode runs a task, cap Brain snapshots, and clean internal reproducible caches.
Recommendation for 1.4.12 users
Update to 1.4.13 as soon as the updater offers it. Do not blindly delete target, node_modules, or build environments unless you know which workflow owns them.
1.5.x: from “file support” toward more consistent “stack understanding.”
The 1.5.x roadmap prioritizes depth, not merely increasing the number of file extensions ChatCode can read.
More semantic parsers
PHP is the first priority, followed by expansion based on real usage into Go, Java/Kotlin, C#, C/C++, Vue/Svelte, and other stacks.
Deeper framework adapters
Laravel will gain deeper understanding; Python web, Node backend, Java/.NET, and popular frameworks will gradually receive dedicated profiles instead of generic runtime handling only.
Typed task runners across ecosystems
The goal is to bring pytest, PHPUnit/Pest, Go test, Maven/Gradle, .NET test, and others into the same safe verification loop as Cargo/package scripts.
Smarter Environment Doctor
Detect toolchains, lockfiles, version constraints, prerequisites, and suitable launch methods per stack; reduce manual configuration.
Storage Doctor & Smart Cleanup
Measure build/cache artifacts, classify them as Safe / Review / Protected, and clean only data proven to be reproducible instead of deleting folders by name.
Capability/plugin architecture
Specialized capabilities can be developed independently and plugged into ChatCode, providing a foundation for faster expansion without making the core increasingly heavy.
What should you use 1.4.x for today?
If your stack is not first-class yet, ChatCode is still useful — you just need the right expectations about automation depth.
Excellent fit
Rust, TypeScript/JavaScript, Node/Vite/Next/Nuxt/Astro/SvelteKit, Laravel/PHP runtime workflows, large projects that need Brain, Git/remote/review.
Useful with more supervision
Python projects without framework adapters, PHP where deeper semantics are needed, Go/Java/.NET/C++, and stacks without typed task profiles yet.
Safety principle
Start read-only, enable Write/Task when needed, let ChatCode report the environment before installing/running anything, and review deploy/destructive actions.