FAQ

General

What is Label Alchemy? Label Alchemy is a VS Code extension that converts hard-coded strings in Salesforce source files (Apex, LWC, Aura) to Custom Labels. It finds the strings, generates valid label names, shows you a diff, and writes the changes when you approve.

Does it work with my Salesforce org type? Yes. Label Alchemy works with any Salesforce DX project structure regardless of org type — Developer Edition, sandbox, scratch org, or production. The deploy feature prevents deploys to production, but scanning and converting work everywhere.

What Salesforce file types does it support?

• Apex classes (.cls)
• LWC JavaScript (.js inside an lwc/ bundle)
• LWC HTML (.html inside an lwc/ bundle)
• Aura JavaScript (.js inside an aura/ bundle)
• Aura markup (.cmp, .app inside an aura/ bundle)

Does it work in the Salesforce Extensions Pack / SFDX workspace? Yes. Label Alchemy is independent of the Salesforce Extensions Pack and works alongside it. Both can be installed at the same time.

What version of VS Code is required? VS Code 1.85 or later.

AI and naming

Do I need an AI/LLM to use Label Alchemy? No. The default naming mode is fully deterministic — it runs locally with no API key, no network call, and no cost. AI naming is opt-in.

How does the deterministic naming work? It splits the string into meaningful words, strips filler words (“a”, “the”, “of”), detects a context hint from surrounding code (error message, placeholder, button, etc.), applies your label prefix if configured, and enforces the 40-character Salesforce name limit.

What AI providers are supported? Anthropic (Claude), OpenAI, Google Gemini, DeepSeek, OpenRouter, Ollama, LM Studio, and any custom OpenAI-compatible endpoint. See the full guide →

What is BYOK? Bring Your Own Key. Label Alchemy doesn’t have its own AI subscription — you provide an API key for whichever provider you use. Your key is stored in VS Code’s encrypted secret storage and used only to call the provider’s API on your behalf.

Can I use AI naming for free? Yes. Per-file AI naming works on the free tier — and if you use Ollama or LM Studio, the model runs locally on your machine at $0. Project-wide AI naming (which reuses your existing labels and follows your naming convention across the whole codebase) requires a license; without one, AI naming falls back to per-file scope.

Can I edit AI-suggested names? Yes. Every name — whether deterministic or AI-suggested — is editable in the diff panel before you approve. The Approve button stays disabled until all names pass validation.

Will the AI ever produce an invalid Salesforce label name? No. Whatever the model returns is sanitized to a valid Salesforce name before it appears in the UI. The model’s raw response can’t reach your codebase in an invalid state.

Privacy and data

Does Label Alchemy send my code to a server? In default (deterministic) mode: no. The scanner, naming engine, and apply step are all local. Nothing is sent anywhere.

In AI naming mode with an API-based provider: the string values from your source files are sent to the LLM provider (Anthropic, OpenAI, etc.) to get name suggestions. File paths and surrounding code context may also be sent.

Use Ollama or LM Studio as your provider if you need to keep everything local even in AI mode.

Is my license key stored securely? Yes. License keys are stored in VS Code’s encrypted secret storage (vscode.SecretStorage), not in settings.json. They are never logged or written to disk in plain text.

Does Label Alchemy have telemetry? No usage telemetry is collected. Label Alchemy does not phone home or report usage statistics. The only external connections are: license validation (Lemon Squeezy) and LLM API calls when AI naming is enabled.

Where is the privacy policy? Privacy policy →

Features and behavior

What happens to strings that appear in multiple files? They map to one label. The second occurrence refers to the same label name instead of creating a duplicate.

Will it overwrite my existing Custom Labels? No. Label Alchemy merges new labels into your existing CustomLabels.labels-meta.xml. It never modifies existing entries.

What if I don’t have a CustomLabels file yet? Label Alchemy creates one at the correct path under your force-app structure.

Why do LWC HTML changes also update the .js file? LWC {binding} syntax in HTML requires a getter in the sibling JavaScript file. Label Alchemy adds the @salesforce/label import and getter automatically so the binding resolves. Without this step, the LWC component would fail to deploy.

Can I organize labels into categories? Yes, free. Set a Salesforce category on labels right in the review panel — per label, or across a whole file in one click. It writes to the label’s <categories>, so labels arrive in your org already sorted.

Does Label Alchemy keep a record of what it changed? Yes — on every convert, no git required. It writes a timestamped folder to labelalchemy-changes/ containing a readable summary.md, a summary.csv, and a portable package.xml you can deploy or hand to a teammate. A one-click Open summary notification appears when the conversion finishes. This is free on every tier.

Can I undo a conversion? Yes — through git. Label Alchemy writes to disk, but it doesn’t commit anything. A git checkout or git restore reverts the changes. We recommend converting on a branch so you can PR-review the diff before merging.

What is the Audit Report? A folder-wide read-only scan that shows every hard-coded string organized by file and component. Free. Use it to understand the scope of work before committing to any conversions.

Does Label Alchemy handle Salesforce namespace prefixes? Generated label references use the default c. namespace ($Label.c.MyLabel, @salesforce/label/c.MyLabel), which works in any org without a managed namespace. Custom managed-package namespaces in references aren’t configurable yet. Separately, the labelAlchemy.labelPrefix setting prepends a string (e.g. MyApp_) to every generated label name for grouping and uniqueness — it does not change the c. reference namespace.

Licensing and pricing

Is the free tier really free? No card required? Yes. Scan, name, review, and apply on single files — no account, no credit card.

What do the Paid features add? Bulk convert across your whole project, deploy to org on approve, project-wide AI naming, a custom denylist, diverging-name duplicate detection, and Audit CSV export — plus priority email support. See the full breakdown in Features.

Is it a subscription? No. The Individual license is a one-time, perpetual purchase — the version you buy is yours to keep forever. It includes a 12-month window of updates; after that, your version keeps working and only newer features lock until you renew updates. A per-seat Team plan is coming soon. For all current prices, the founding-member offer, and renewal costs, see labelalchemy.dev/#pricing.

Is there regional pricing for India or other markets? Yes. Lemon Squeezy’s Purchasing Power Parity pricing applies automatically at checkout for India and other markets. No code needed.

Is the license per machine or per user? It activates a limited number of devices (seats). You can move it yourself with Label Alchemy: Deactivate License on This Device on the old machine, then re-enter the key on the new one — see License activation.

What happens if my internet is down and license validation fails? Label Alchemy has a 3-day grace period. If it validated successfully before and can’t reach the license server now, you continue working normally. A warning is shown after 1 day.

What’s on the roadmap? Directional next-ups include a PR / git-diff CI scan, deeper project-context AI naming, and async batch deploy. See the public roadmap →.

Can I use the extension in CI? The extension runs in the VS Code UI context, not as a standalone CLI. It’s designed for interactive developer use, not CI pipelines.

Troubleshooting

Common problems and fixes — scanner found nothing, package-root errors, capability-test failures, blocked deploys, license-key issues — are on the dedicated Troubleshooting page.