Research Desk / receipt 01

No magic
numbers.

Every number below names the host, workload, source commit, and limitation. Re-run the same script instead of trusting a floating performance claim.

node scripts/benchmark_research_desk.mjs

Loading the checked-in receipt…

One host · five disclosed measurements

The current receipt.

Use these values to understand the tested build, not to predict every machine. Raw samples remain in the public JSON.

01 / Package—

Uncompressed native application bundle.

02 / Native init—

Median fresh-data embedded-runtime smoke start.

03 / Peak memory—

Median maximum resident set during the same starts.

04 / Index parser—

Exact app parser and fingerprint path; native I/O excluded.

05 / Warm build—

Median repeated production Vite build, including npm startup.

Host—

CPU—

Runtime—

Commit—

Collected—

Read the fine print first

The receipt intentionally separates observed product behavior from what still needs a native end-to-end measurement.