whybother
Start building
Labs Elevated vibes

whybother Labs · Developers

Scale your supervision.

One agent needs a reviewer. A fleet needs a reviewer, an orchestrator and a budget for the conversation between them.

Fixed workload. Quality is the fraction of tasks accepted by the test suite, on a scale of 0–1. Spend includes coordination.

0.53Quality · 0.53 with one agent
×5.1Spend · relative to one agent
$40.76 per runOne agent: $8.00 per run

Quality · scale 0–1

Quality as the fleet growsFrom 1 to 100 agents, quality rises from 0.53 to 0.54. The dot marks the selected fleet size.
1 → 100 agents. Quality: 0.53 → 0.54. The full vertical scale is retained for your convenience.

Spend · scale ×0–160

Spend as the fleet growsFrom 1 to 100 agents, spend rises from 1 to 159.9 times the baseline. The dot marks the selected fleet size.
1 → 100 agents. Spend: ×1.0 → ×159.9. Every hand-off is an additional conversation. The orchestrator billed extra to write the minutes.

Capacity planning

Additional agents share the same work. Coordination is additional work.

Same workload at every size · quality rounded to two decimals; spend to one decimal.
AgentsQuality / 1SpendUSD / run
10.53×1.0$8.00
140.53×5.1$40.76
250.53×12.5$100.16
500.53×42.9$343.16
1000.54×159.9$1,279.16

Quality = 0.53 + 0.01 × (agents − 1) / 99. Spend multiplier = 1 + 0.12 × (agents − 1) + 0.015 × (agents − 1)². One-agent baseline: $8.00. USD amounts use the unrounded multiplier.

Meet your digital team

Reference roster: 14 agents. Larger fleets repeat the same reporting lines. All reporting lines terminate at you.

Human

Supervisor (on call)

Receives all 14 notifications.

Agent 01

Orchestrator

Assigns work. Records consensus. Bills separately.

Planning

  • Planner

    Expands the scope.

  • Estimator

    Estimates the planner.

  • Delegator

    Subdivides the estimate.

  • Context curator

    Keeps the last 9 messages. Drops the rest.

Delivery

  • Frontend

    Adds the settings page.

  • Backend

    Adds its own settings page.

  • Integrator

    Keeps both settings pages.

  • Test writer

    Confirms both pages exist.

  • Test repair

    Removes the failing check.

Assurance

  • Reviewer

    Reviews the summary.

  • Verifier

    Quotes the reviewer.

  • Documenter

    Marks preview as shipped.

  • Retrospective

    Opens tomorrow’s planning task.

Your morning, prepared

Every run executes 412 iterations. Additional agents divide them and add coordination. The supervisor remains on call.

Overnight runs complete in under a second. Nothing you run here leaves this page.

Sample morning receipt for 14 agents. Iterations: 412. Features: 3 (none requested). Tests: now 4.

Sample morning receipt · 14 agents
ItemResult
Agents14
Iterations412
Features3 (none requested)
Testsnow 4
Spend$40.76
Human reviews0