QLRO
For Braket spenders

You spent money on AWS Braket last month.
Here's what Qlro would have saved you.

Paste your AWS credentials? No. Run one command locally. It lists every completed Braket task you ran in the last 30 days, backfills what Qlro would have predicted for each, flags tasks that would have been cheaper on a higher-fidelity alternative, and gives you a public link with the numbers.

One command

pip install "qlro[braket]"
qlro braket-retro --days 30

Uses your local AWS credentials via the Braket SDK — nothing leaves your machine until you see the summary, and even then only circuit shape and observed fidelity are submitted (no gate parameters, no measurement outcomes beyond the aggregate). The CLI prints a shareable URL like https://qlro.io/predict/braket-retrospective/retro_xxxx.

Observability
Every task ran on the device you happened to pick. Qlro retroactively computes what it would have predicted and where reality matched vs. diverged.
Cost visibility
Fidelity-per-dollar ranking on 4 vendor backends. If a cheaper device would have kept fidelity above the threshold, we show the delta.
Accuracy dashboard
Every (predicted, observed) pair flushes into qlro.io/accuracy. Your session adds to the public cross-vendor ground truth Qlro publishes.

Pricing

Free tier: 2 retrospective sessions per week per IP. No account. No credit card. The public session URL stays live — cite it from a paper, share it with a collaborator, embed it in a weekly report. For higher frequency, email official@stockfolio.ai.

Server: https://qlro-production.up.railway.app/api/v1/retrospective/braket. CLI source in qlro.cli.