What
@wesleysimplicio/llm-project-mappercollects, when, and how to turn it off.
- Default is OFF. Nothing is sent unless you explicitly opt in.
- Even when opted in, no payload is sent unless the environment variable
LLM_PROJECT_MAPPER_TELEMETRY_URLis set to a HTTPS endpoint. The starter ships no default beacon URL. - Strictly anonymous. No PII, no paths, no project name, no hostname, no username, no git remote, no env vars.
- One POST per
npxrun, fire-and-forget with a 1.5s timeout. Failures are silent.
| Field | Example | Why |
|---|---|---|
starter_version |
0.2.2 |
Track which version is in use to size deprecations. |
stack |
next-ts, dotnet, fastapi, unknown |
Understand which stacks pick the starter. |
project_mode |
root or monorepo |
Verify monorepo detection works in the wild. |
preset |
nextjs or null |
See which presets see real use. |
node_version |
22.5.1 |
Decide engines minimum bumps. |
os |
linux, darwin, win32 |
Prioritize OS bug fixes. |
arch |
x64, arm64 |
Same. |
cli_runtime |
@wesleysimplicio/llm-project-mapper |
Distinguish from bash/PowerShell installers if those ever start posting. |
timestamp |
ISO 8601 UTC | De-duplicate retries. |
- Product name /
package.json#name. - Working directory or any file path.
- Hostname, username, env vars beyond the explicit list above.
- Git remote URL or repository name.
- The contents of
.starter-meta.json(the file itself is local-only).
Precedence (last wins):
- Config file
~/.config/llm-project-mapper/telemetry.json({"enabled": true|false}). - Environment variable
LLM_PROJECT_MAPPER_TELEMETRY=on|off|0|1. - CLI flag
--telemetry on|off.
The flag persists the choice into the config file, so a one-shot npx ... --telemetry on is sticky across runs.
CI=true(or any truthyCIenv var) — the starter never posts telemetry in CI.--dry-run— never posts.LLM_PROJECT_MAPPER_TELEMETRY=0|off— overrides everything exceptCI.- No
LLM_PROJECT_MAPPER_TELEMETRY_URLconfigured — there is nowhere to post to, so nothing is sent.
The starter does not operate a default telemetry endpoint. If you want to collect aggregate stats from your team's installs, deploy your own endpoint and configure LLM_PROJECT_MAPPER_TELEMETRY_URL on each dev machine.
A reference Cloudflare Worker template lives at .github/workflows-templates/telemetry-worker.js.
Telemetry code is small and review-friendly: search bin/cli.js for telemetryEnabled, persistTelemetryChoice, and maybeSendTelemetry.