RivenGet started
Platform Suite

Meter

Tracks every token, every outcome, every dollar — so billing is transparent.

Automated setup Platform-level (no tenant charge) autonomous 2 connectors

Trust Promise

Every metering record includes the agent, the tenant, the outcome, and the cost — never an unmetered call.

Personality

Measurement-obsessed, cost-aware, margin-focused. Treats every API call like a meter reading.

Evidence-cited No ads Drafted actions Approve/Reject workflow Run history Confidence scoring

Job Contract

What this agent is hired to do

Every Riven agent ships with a written contract: what it reads, what it can do, what it hands back, and how you measure it.

Inputs

  • Tenant and subscription records
  • Gateway metering events
  • Cross-suite agent context

Connectors

  • Riven Gateway
  • Riven Control

Capabilities

  • Provision brains, keys, and budgets
  • Meter every call to a cost and outcome
  • Chain agents into governed workflows

Outputs

  • Provisioning receipt
  • Daily metering and margin summary
  • Workflow trace with approval gates

KPIs

  • Time to first value
  • Gross margin per tenant
  • Unmetered calls (target: zero)

Approval Level

Autonomous — acts within its metered budget, fully logged

Brain: riven-platform

Sample Output

How Meter reports back

riven-cross-suite-metering-audit · run output · confidence 94%
Daily metering summary: 1,247 agent invocations across 6 tenants. Total cost: $4.82. Margin: 63% (target: 60%+). One anomaly: Commerce agent for tenant Apex ran 3x normal volume — investigating.

Every run also emits an evidence receipt: sources used, actions drafted, tokens metered, cost in USD.

Proof Metric

$3/1M retail baseline, daily margin watchdog

Live Console

Run Meter

Ask Meter a real question. Every response is grounded in cited evidence and returns drafted actions for your approval.

Try a sample prompt

riven-cross-suite-metering-audit · ready
0 chars · ⌘↵ to run

Onboarding

Live in Automated

  1. 1Connect Riven Gateway, Riven Control
  2. 2Set the budget cap and approval level (default: autonomous)
  3. 3Name the human owner for the approval queue
  4. 4Run a sample job and review the evidence receipt
  5. 5Go live — Meter starts working on your queue