FAQ

Do I need to send events, or is connecting my cloud enough?

Connecting your cloud is enough for most customers — it gives cost and usage attribution from your provider's own billing and logs. Add events only if you want per-request resolution (tokens, cost, latency, identity per call).

Is there an SDK?

No Venturi-specific SDK is required. Integration is the read-only Terraform connector plus, optionally, a plain HTTP ingestion API or a drop-in proxy you point your existing provider SDK at. To read attributed cost, energy, and carbon back out programmatically, use the REST API, which is backed by typed TypeScript and Python SDKs.

Can Venturi change anything in my account?

No. Every onboarding module is read-only, a platform gate blocks any module that requests a write from shipping, and the bundled verify.sh proves writes fail. See Trust & security.

Does Venturi see my prompts or completions?

No. The pipeline never stores prompt or completion text; the event schema has no content field. Only metadata (tokens, model, cost, identity, timing).

Will the proxy slow down or break my traffic?

No. The interceptor is fail-open on a hard latency budget — if Venturi is slow or down, your request is forwarded to the provider anyway.

Where does Venturi run?

In a dedicated data plane inside your cloud trust boundary, so your operational data stays with you. See How it works.

How do I rotate or revoke access?

Rotate with terraform apply (new external_id / federation subject), revoke with terraform destroy. Details in Rotate & offboard. Tell Venturi first so ingestion pauses cleanly.

How long does onboarding take?

~15–30 minutes of hands-on work. Venturi confirms ingestion within one business day, after which the dashboard populates in ~1–2 hours.

What's attribution_confidence and why is it capped?

Every attribution carries a confidence score, capped at 0.95 — Venturi never claims certainty. The cap is enforced in the schema and the scoring engine.

Which providers are supported?

Cloud-native inference (Bedrock, Vertex AI, Azure OpenAI) via the connector; direct OpenAI and Anthropic via the proxy or ingestion API. Ask your onboarding contact about others.

How does Venturi know how much energy and carbon my AI usage produces?

For every invocation it can identify, Venturi estimates GPU energy from the token counts, the model identity, and a per-model energy coefficient drawn from the AI Energy Score — a standardized, independent benchmark maintained by Hugging Face, Salesforce, Cohere, and Carnegie Mellon University. Carbon is then that energy multiplied by your deployment region's grid intensity. The full formulas, constants, and provenance are in the energy and carbon methodology; the customer-facing summary is on the Energy & carbon page.

Why do some models show no energy or carbon number?

Because the model is not in the energy catalog, and Venturi reports null, never zero. A zero would falsely imply the model consumed no energy; a null honestly says Venturi does not have a coefficient for it. The same honest-unknown rule applies to water and the eco-efficiency score — any unknown input yields null, not a fabricated number. See the honest-unknown rule and the methodology.

How accurate are the carbon numbers for my region?

Carbon is energy × grid intensity, and Venturi resolves grid intensity by a fixed order of precedence — an explicit value you supply, then a live feed, then your cloud region, then your country's generation mix, and only as a last resort the world-average fallback of 475 gCO2e/kWh (always flagged as estimated). Configuring your actual deployment region moves you up that chain and makes the numbers materially more accurate than the fallback. The full resolution chain and fuel-factor table are in the methodology.

Can I set energy budgets, and is energy the same as cost?

You can set per-team and per-cost-center energy budgets (kWh per period), evaluated with the same advisory and gate semantics as cost budgets — see budgets and alerts. But energy is not cost. Energy is sustainability context; an energy multiplier must never be read as a cost multiplier. Reasoning modes, for instance, can be hundreds of times more energy-intensive than standard operation, which says nothing about price. Cost comes from the billing and pricing path; energy comes from the energy engine. The two are always reported separately.

How long is my data kept, and can it be erased?

Venturi stores only metadata — never prompt or completion content — under a defined retention schedule, and supports data-subject erasure requests. See data privacy & retention and data-subject rights.

How do I read attribution data out programmatically?

Use the read REST API: GET /api/v1/attribution returns attributed cost, energy, and carbon with an evidence card on every record. It is backed by typed TypeScript and Python SDKs. Note your base URL depends on your deployment mode — SaaS tier uses https://api.venturi.systems; self-hosted uses your own instance host.

Where do I send feedback or feature requests?

Through the feedback portal at feedback.venturi.systems, or via your onboarding contact. For support questions, see Support.

Can Venturi show an individual employee's AI usage?

No. Adoption and cohort surfaces enforce k=5 suppression — any view that would isolate a cohort of fewer than five people returns the parent rollup, never an individual. Venturi is built to report on teams and cost centers, not to surveil individuals. See tenant isolation and data privacy & retention.

Who do I contact?

[email protected] — see Support.