SideNote ProSpeed and cost
Make answers arrive sooner and keep what you spend on your own provider predictable.
The real contributors to latency, and which ones you control.
A three-profile setup that makes switching a single click.
A test for deciding when depth is worth the wait.
Worked setups for a named model
A complete configuration from start to finish, with what it costs.
A ten-minute, five-dollar DeepSeek setup, start to finish.
Set up GPT-6 Luna for fast everyday answers at a lower API cost than GPT-5.6 Luna.
Set up GPT-5.6 Luna for remarkably fast everyday answers at a tiny API cost.
Working effectively
Habits that make a docked assistant part of the day rather than a novelty.
Configuration that makes asking a question cheaper than switching windows.
Write project facts and templates down once, and attach only the notes a request needs.
Habits that keep a growing archive findable.
Formulas, error values and cell text, answered beside the sheet instead of in a browser tab.
A two-profile split that makes the private choice the easy one.
Presets for the ordinary case, your own endpoint for everything else.
Workflows by profession
How the same panel is used on a developer's, product manager's, consultant's or writer's day.
Repository questions, error triage, and knowing when to escalate.
Ambiguity hunting, proposal comparison, and documentation queries.
Document comparison, decision matrices, and defensible output.
Rewriting in place, cheap variants, and reviewing your own drafts.