Official VS Code extension for Vitte. Vitte Studio gives you a complete daily workflow in VS Code: edit, navigate, diagnose, build, run, test, and debug from one place. This README is intentionally practical. It focuses on what you can do right now, why each command matters, and how to troubleshoot quickly when something breaks.
Vitte Studio supports both .vitte and .vit files.
You get syntax highlighting, autocomplete, hover details, diagnostics, symbol navigation, rename, formatting, semantic tokens, debug integration, and project commands.
In short: less manual setup, faster iteration, and better code feedback while you type.
The extension adds a dedicated Vitte sidebar.
Vitte Documentation loads official docs from https://vitte.netlify.app.
Vitte Summary gives chapter-based navigation from beginner to professional topics.
Vitte Diagnostics aggregates workspace issues.
Vitte Modules helps you inspect project structure.
Vitte Metrics exposes runtime/server metrics useful for performance checks.
Vitte Offline explains offline mode and gives recovery actions.
Open Extensions in VS Code, search Vitte Language Support, then click Install.
- Open VS Code.
- Run
Extensions: Install from VSIX.... - Select your
.vsixfile. After installation, open a.vitteor.vitfile to verify language features are active.
Step 1: Run Vitte: Detect Toolchain.
This checks that the expected binaries are reachable.
Step 2: Run Vitte: Open Docs.
Keep docs visible while coding to reduce context switching.
Step 3: Run Vitte: Quick.
Use it as your daily launcher for build/run/test routines.
Step 4: If features feel stale, run Vitte: Restart Language Server.
Vitte: Detect Toolchain validates runtime/build/lsp paths and helps avoid hidden config drift.
Vitte: Open Docs opens the official documentation quickly, useful during syntax or API checks.
Vitte: Open Playground gives a safe space for quick experiments.
Vitte: Quick opens a curated action menu that saves time during repetitive workflows.
Vitte: Build compiles your project and confirms build health.
Vitte: Run builds and executes your app using current settings.
Vitte: Test runs the test suite for confidence before commit.
Vitte: Restart Language Server resets IntelliSense when completion/hover/diagnostics degrade.
Vitte: Show Server Log opens technical logs for root-cause analysis.
Vitte: Show Server Metrics helps identify slow requests and server pressure.
Vitte: Refresh Diagnostics forces diagnostic refresh.
Vitte: Open Diagnostic jumps directly to a selected issue.
Vitte: Refresh Structure reloads structural/module data.
Vitte: Run Benchmarks and Vitte: Run Benchmarks and Open Report support performance testing loops.
Vitte: Open Bench Directory opens report output quickly.
Vitte: Clean removes build artifacts when you need a fresh state.
Vitte: Test Current File runs file-focused tests for tighter feedback.
Vitte: Switch Build Profile swaps between dev, test, release, and bench.
Vitte: Toggle Incremental Build controls faster rebuild strategy.
Vitte: Toggle Telemetry and Vitte: Quick Export support local observability workflows.
For module-focused work in this repository, use workspace tasks from .vscode/tasks.json:
vitte: parse modules (current file)vitte: strict modules (current file)vitte: module graph (current file)vitte: module graph json (current file)vitte: module doctor (current file)vitte: dump module index (current file)vitte: modules testsvitte: modules snapshotsvitte: module shape policyvitte: runtime matrix modulesvitte: ci mod fast
These map directly to Vitte module tooling (mod graph, mod doctor, strict module checks, and module CI targets).
If auto-detection is enough, keep defaults.
If not, configure these first:
vitte.toolchain.rootvitte.runtime.pathvitte.lsp.pathvitte.build.pathvitte.fmt.path
Then tune behavior:
vitte.build.profilevitte.build.incrementalvitte.bench.reportDirvitte.telemetry.enabledvitte.telemetry.sampleRatevitte.quickActions.sequencesvitte.quickActions.defaultsvitte.commandShortcuts
Start with Quick to run common actions.
Use Build after major edits and Test before pushing changes.
Use docs and summary views while reviewing unfamiliar syntax.
When behavior looks inconsistent, restart the language server before changing settings.
If autocomplete disappears:
- Run
Vitte: Restart Language Server. - Check
Vitte: Show Server Log.
If toolchain detection fails:
- Run
Vitte: Detect Toolchainagain. - Set explicit paths in settings.
If extension enters offline mode:
- Run
Vitte: Explain Offline. - Open
Vitte: Open Offline Log. - Generate
Vitte: Copy Offline Report.
The extension provides debug type vitte.
Typical scenarios include launching the current file, launching project entry, and attaching to a running session.
The extension keeps one explicit activation event: onStartupFinished.
Command, view, and language activation are inferred by VS Code from contributes in package.json.
Version history: CHANGELOG.md.
License: MIT.
The extension now ships a powerful suggestions stack with safe defaults.
-
Local inline engine (default path):
- repo-local indexing (
.vit/.vitte/.vitl) - n-gram + AST-lite patterns
- scope-aware, flow-aware and expected-value hints
- adaptive learning from accepted suggestions
- hot-context LRU cache for low latency
- repo-local indexing (
-
Optional AI pipeline:
- local RAG top-k chunks
- optional dedicated backend endpoint
- reranking + dedupe + syntax post-processing
- anti-hallucination filtering against known symbols
- enriched context from diagnostics/build/test/git/dependencies
- Cloud opt-in: off
- Local-only mode: on
- Trusted workspace only: on
- Secret redaction: on
- External training declaration: off
Use:
Vitte: Suggestions Cloud Opt-InVitte: Suggestions Cloud Opt-Out (Local-Only)
Vitte: Inline ToggleVitte: Suggestions Refresh ContextVitte: Suggestions Open ProfilerVitte: Suggestions Show Local Engine StatsVitte: Suggestions Profile ResourcesVitte: Suggestions Export Diagnostics
vitte.suggestions.granularity:classic_only|inline_only|hybridvitte.suggestions.localEngine.*(local ranking, cache, multiline, adaptive learning)vitte.suggestions.aiPipeline.*(backend, RAG, privacy, worker indexing, prompt cache)