GVGanivra

AI unit economics

Know which AI customers and workflows make money.

AI usage can grow while profit disappears. Connect customer revenue to recorded model and tool costs, find margin leaks, and fix them before you scale the losses.

See contribution by customer, feature, and workflow—then inspect the executions driving the difference.

Start free trial

14 days · No credit card required

Check your AI economics
02 lines to install↗ no proxy∅ no prompts stored✓ fail-open telemetryAutomatic + explicit coverage · OpenAI · Claude · Gemini · Perplexity · MCP

The cost behind every result

Customers pay for the result. You pay for every attempt.

More accurate extraction, better answers, and more capable agents can require extra validation, longer context, and repeated attempts. Your customer price may stay the same while the cost of delivering each result grows.

Validation passesLonger contextAgentic reviewPremium modelsRetries
Illustrative customer$50 / month plan
At launchAI cost-to-serve
$10+$40 contribution
Six months laterAI cost-to-serve
$80−$30 contribution

The provider bill shows $80 of spend. Ganivra shows which customer caused it, which workflow changed, and what to do next.

01Which customers are profitable?
02Which workflows cost more than their price covers?
03What changed when delivery costs increased?
04Which customers may fall below our margin target?

Follow the economics

From a customer’s margin to the calls behind it.

  1. Customer
  2. Product / feature
  3. Workflow / agent
  4. Execution
  5. Model / tool costs

Customer C · illustrative month

Where did the money go?

$610recorded AI cost
Research workflow$380
62% of recorded cost
Report generation$160
Other workflows$70

Start with research. Inspect its model calls, tools and reported retries.

From insight to action

See the driver.
Follow the evidence.

  1. 01
    Find the customer

    Connect revenue to the cost of serving them.

  2. 02
    Isolate the workflow

    See which features and agents drive recorded spend.

  3. 03
    Inspect the execution

    Follow the model and tool calls behind the total.

Connect your first workflow →
What these numbers include

Revenue may be reported, plan-rated or allocated. Missing prices and revenue stay visible. Retrieval, telephony and compute require instrumentation and pricing. Each recorded call contributes once.

MCP metering and cost attribution

The tool call is only one line on the trace. The cost can keep growing after it returns.

MCP gives agents access to data and actions. It also creates hidden cost paths: a single tool decision can trigger retrieval, follow-up generation, validation, and retries. Instrument the model and tool calls with a shared execution ID. Ganivra joins those recorded steps so you can investigate the cost of the workflow.

See cost by server + toolTrace downstream model callsCatch reported loops and retriesPrice the full customer action
One customer request$0.212 total
  1. 01
    Agent plansGPT-5
    $0.040
  2. 02
    search_accountsMCP · crm-server
    $0.002
  3. 03
    Agent summarizesGPT-5
    $0.110
  4. 04
    Validation retryGPT-5 mini
    $0.060

0.2¢ direct tool cost became 21.2¢ of execution COGS. The protocol is not the problem; invisible fan-out is.

Agentic workflow economics

One result. Every extra step adds up.

Extraction is only the start. Validation, review and retries can turn a simple document task into a much more expensive execution.

See the recorded cost of the complete workflow before changing your pricing or implementation.

Extract→Validate→Review→Deliver
Explore agentic workflow economics →
Illustrative document workflow · tracked AI costs
Standard processing$0.09
Processing with review + retries$0.31
Same document. More work.3.4× the costSee which recorded steps contributed to the increase before changing the workflow.
Unit economics snapshotIllustrative · with revenue metadata

AI revenue

$6,420Example revenue metadata

AI cost

$1,842.67Model + MCP-linked COGS

Contribution margin

71.3%$4,577 after recorded AI cost

Cost / execution

$0.048Workspace average

Below target margin

3Illustrative customers

Prompt-level unit economics

Know which prompt versions make or lose money.

Illustrative product direction · with revenue metadata
Margin regressionticket_resolution / prompt:v17
+28%cost per execution

Prompt v17 added more context and longer outputs. Cost rose from 4.8¢ to 6.1¢ while contribution margin fell eight points.

Prompt versionCost / executionContribution marginChange
v17Current$0.06163%−8 pts
v16Baseline$0.04871%—
v15Previous$0.05268%−3 pts

No prompt content stored. Ganivra attributes cost using customer-supplied prompt_id and prompt_version metadata.

The profitability gap

Your AI bill cannot tell you whether your AI product is profitable.

Provider spend is a total. Profitability requires customer context.

Provider bill saysGanivra reveals
GPT-5 spend: $1,054Prompt v17 raised cost 28% and cut margin 8 points
Total AI cost: $1,843Northstar Health costs 2.7× average
Requests ↑ 18%Reported retry behavior adds $11K/month exposure
1 MCP tool call3 downstream model calls drove 99% of execution cost
429M tokensThree customers fell below target margin
What you spentWhere margin is won or lost

Customers. Workflows. Agents.

See what it costs to deliver each result.

Model + MCP events · one API
01 · Outbound calls

Is your calling agent earning its keep?

Connect each customer’s qualification and follow-up calls to the agent’s recorded model and tool costs. Compare those costs with revenue to see what’s left.

Connect a calling workflow →
02 · Document processing

Same price per document. Different costs.

Follow extraction, validation, and review agents through one customer’s document workflow. Find where extra steps consume the revenue you earn per document.

Explore document workflows →
03 · Research agents

What did that finished report really cost?

Group a customer’s research, retrieval, and writing steps into one execution. See which agents drive recorded spend before setting a price for the workflow.

Explore research workflows →
04 · Coding agents & DevTools

Inspect expensive agent runs

Follow model steps, token usage, and reported retries through a coding workflow. Find where repeated generation and testing add cost.

Explore coding agents →
05 · Multi-tenant B2B SaaS

Know your cost per customer

Use existing tenant IDs to compare usage, revenue, and margin. See where customer pricing or usage limits need a closer look.

Explore B2B SaaS →
06 · Regulated industries

Share usage facts, not sensitive content

Send model names, tokens, latency, and pseudonymous IDs. No prompts, responses, PHI, or customer contact details required. No Ganivra proxy.

Explore regulated enterprises →

AI usage metering by customer and workflow

Where cost-to-serve accumulates

Illustrative B2B data · metadata only

By model

Cost
gpt-5$1,054.80
gpt-5-mini$438.52
gpt-4.1$244.10
gpt-4.1-mini$105.25

By MCP tool · linked COGS

Cost
search_accounts$388.12
retrieve_policy$274.45
update_ticket$189.90
check_availability$96.22

By workflow

Cost
inbound_support_call$741.28
ticket_resolution$506.81
document_processing$367.43
sales_follow_up$227.15

By feature

Cost
voice_agent$698.02
support_copilot$467.36
workflow_automation$401.12
ticket_triage$276.17

By account

Cost
Northstar Health$315.48
Atlas Logistics$248.91
Meridian Financial$196.20
Summit Retail$164.87

Margin risk queue

Executions eroding margin

5 shown
ExecutionWorkflowAccountCallsTokensCostLatencyTimestamp
exec_demo_72f1inbound_support_callNorthstar Health422K$0.317.20sAug 07, 2026 · 12:00 UTC
exec_demo_91acticket_resolutionAtlas Logistics529K$0.28411.92sAug 07, 2026 · 11:42 UTC
exec_demo_0bd4document_processingMeridian Financial422K$0.2478.64sAug 07, 2026 · 11:18 UTC
exec_demo_c822sales_follow_upSummit Retail619K$0.21914.32sAug 07, 2026 · 10:51 UTC
exec_demo_f10einbound_support_callNorthstar Health39.7K$0.1835.84sAug 07, 2026 · 10:37 UTC

From telemetry to financial evidence

Keep recorded cost and revenue actuals connected.

Report confirmed purchases, subscriptions, credit purchases and refunds. Add telephony, speech, OCR, compute and other delivery costs; reconcile provider bills, review budgets, save evidence and retain administrative history. Ganivra records these facts—it does not process payments or manage customer balances.

Explore the integration guide →

Frequently asked questions

What teams ask before instrumenting.

Straight answers about data, integrations, and getting started.

01What is Ganivra?

Ganivra is an AI unit economics platform that shows which customers and workflows make money. It connects customer revenue to recorded model, MCP tool, and delivery costs so teams can find margin leaks, improve pricing, and avoid scaling unprofitable usage.

02What is AI metering?

AI metering records the resources an AI application consumes, such as tokens, model calls, and tool invocations. Ganivra attributes that usage to customers, features, workflows, and agents so teams can understand cost-to-serve and evaluate their pricing.

03Is AI metering the same as usage-based billing?

Metering measures consumption; billing applies customer prices, allowances, and contract terms to calculate charges. Ganivra provides usage and cost attribution to inform pricing decisions. Customer invoicing and real-time usage enforcement require separate billing and application logic.

04How does Ganivra calculate AI unit economics?

Ganivra attributes instrumented model calls, MCP tool steps, and reported retries to product metadata such as customer, feature, workflow, agent, prompt ID, server, and tool. Revenue can be reported, calculated from a configured customer pricing plan, or allocated across executions. Contribution is revenue minus recorded AI costs; it is not net profit or full gross margin. Missing revenue and unpriced costs remain visible.

05Why does MCP cost tracking matter?

An MCP tool call may be free or inexpensive by itself, but it can trigger planning, retrieval, validation, retries, paid APIs, and additional model calls. Ganivra joins instrumented steps under one execution so teams can see the recorded cost associated with each server and tool instead of only the direct tool fee.

06Can I track outbound calls, document processing, and research agents?

Yes. Send customer, workflow, and agent labels with your model and tool events, and use a shared execution ID for steps belonging to one customer action. Connect revenue or a configured pricing plan to compare revenue with recorded costs. Workflow labels do not automatically capture telephony, OCR, retrieval, or compute costs; those require separate instrumentation and pricing.

07Does Ganivra store prompts or responses?

No. Ganivra is designed to collect execution metadata rather than prompt or response bodies. Prompt-level economics use customer-supplied prompt IDs and version labels.

08Which AI providers does Ganivra support?

Python SDK 0.6.1 automatically captures synchronous, non-streaming OpenAI Responses and Claude Messages sync, async, and streaming with the Anthropic extra. It can also record MCP tools with ganivra.mcp_call. Gemini, Perplexity, other APIs, and unsupported provider modes use the explicit SDK client or REST event contract. Unknown costs remain explicitly unpriced until a catalog, contract, or provider-reported rate is available.

09How do I get started?

Start a 14-day full-product trial, copy your SDK key, and instrument one real workflow. Your first execution will appear in the dashboard with attributed cost.

Learn how to measure consumption in our AI metering guide, then compare token, credit, and outcome-based billing.

Start your full trial

Your first workflow is enough to start finding value.

Start with one workflow. See which customers create margin, which consume it, and what must change before your next growth push or investor conversation. No sales call or credit card required.

01 See the cost of your first recorded execution02 Find expensive customers and workflows03 Connect revenue to compare cost and contribution
Initialize the Python SDK in two linesimport ganivra
ganivra.init(api_key="your_key")
Start with one workflow

See your first real execution

No credit card
01Create a workspace
02Copy your workspace API key
03Connect via SDK or REST and send usage
Start free 14-day trial

Already have an account? Sign in

Python SDK · REST APIOpenAI · Gemini · Claude · Sonar · MCPMetadata only · Fail-open