Teams

Calx scales
from one supervisor to an org.

Same compiler. Different operator surface. The feature set grows as your team does. Pricing lives in the audit pipeline: book the audit, we scope the deployment, everyone starts with the same proof.

View

Every deployment starts with a Correction Audit. Pricing scoped per engagement.

Individual · one operator

One person. One harness. Corrections that compound.

For technical founders and independent practitioners dogfooding agents daily. Correction capture, recurrence clustering, and runtime enforcement for a single operator. Your rules stop being your job.

Who uses it

Solo practitioners, dogfooding founders, first engineers on a rollout

Available via audit

Core compiler

●
Bench · correction capture
Inline correction, scoped to operator identity.
●
Recurrence detection
Cluster detection across sessions.
●
Tether · compiled rules
User-promoted rules run as middleware hooks.
●
Evidence ledger
Prevented-violation log per rule, per session.

Scope

●
1 operator
Identity-scoped rule compilation and enforcement.
●
Unlimited rules once compiled
Free-tier class has a compiled-rule ceiling; paid individual removes it.
●
BYOK or managed inference
Your model provider, your keys, or Calx-managed.
●
macOS desktop · Linux beta
Bench ships on desktop. Web surface follows.

Start with a founder-to-founder audit. Spencer reviews your correction workflow, classifies the recurrences, and ships a 2-week enforcement plan. Individual tier unlocks after.

Book an audit →

Full feature comparison

All tiers, side by side.

Starting tier is Free: one operator, 25 compiled rules, BYOK. Above that, every engagement begins with an audit and a scoping call.

Feature set by tier

No prices shown · scoped per engagement