The Plan Sets What a Run May Spend

neatlogs measures every doc-test run. cfo.ai turns those costs into tier prices and a cost cap per run. FirstRun stops a run at that cap, and a neatlogs alert fires when real runs drift past it.

  1. neatlogs measures1
    Every run's cost

    Read from the firstrun.run traces.

  2. cfo.ai plans2
    Prices and caps

    18 months, 6 Scenarios.

    Open the model
  3. FirstRun enforces3
    A cap per run

    firstrun run --tier scale

  4. neatlogs watches4
    An alert on drift

    p95 run cost above the Scale cap.

Each Tier Gets a Cap That Keeps a 70% Margin

Scroll sideways to see every column.

TierPriceRuns a monthCap per runRuns the cap stopsMargin at the mean runMargin at the p95 run

How FirstRun Makes Money

  • The buyer is a DevRel or docs team at a dev-tool company.
  • Teams pay a monthly tier: Free $0, Team $49, Scale $299.
  • Plan, no paying customers yet.

Measured Cost per Run Against the Caps

Each bar counts published doc-test runs in a cost band.

Cash on Hand Through April 2028

cfo.ai Scenarios with the same customers. Only the cap or the Scale price changes.

The Guardrails Fired 3 Times in 2 Days

    neatlogs alert history, read with the CLI

    Refresh it with firstrun econ, then python scripts/plan_snapshot.py. Model costs are list price, not billed.