Frequently Asked Questions

Tailpipe is a platform that accurately measures and helps reduce the carbon emissions from your cloud infrastructure and AI workloads. We provide real-time data from your actual usage — not estimates or industry averages.

Tailpipe supports AWS, Microsoft Azure, and Google Cloud Platform (GCP). We provide comprehensive coverage across compute, storage, networking, and AI services.

We use your actual cloud usage data combined with real-time grid carbon intensity data to calculate emissions. Our methodology is aligned with the GHG Protocol and verified by independent experts.

Tailpipe supports GHG Protocol Scope 2 and 3 reporting, and is aligned with the Green Software Foundation SCI and SCI-AI specifications. Reports are audit-ready for compliance purposes.

Most customers are up and running within a day. Our baseline assessment covers 3 months of historical data, giving you immediate insight into your cloud carbon footprint.

No. Tailpipe uses lightweight, open-source scripts to configure read-only access to your cloud billing and usage data. There are no agents, daemons, or software packages to install in your cloud environment.

No. Our setup scripts create only the IAM roles and policies needed for read-only data access. They do not modify, create, or delete any of your existing cloud resources.

You'll see your first data within 24 hours of completing the integration. Our baseline assessment then covers 3 months of historical data, giving you immediate insight into trends.

Yes. The integration is fully reversible. You can remove the IAM roles and policies created by our setup scripts at any time, and we will delete all collected data on request.

Completely. The setup scripts create clearly-labelled resources (prefixed with 'tailpipe-') that can be removed in a single step. We also provide removal scripts.

Yes. Tailpipe supports multi-account setups across all providers, including AWS Organizations. Each account is connected separately with its own read-only role.

Tailpipe requires read-only access to your cloud billing and usage data. For AWS, this means the Cost and Usage Report and CloudWatch metrics. We follow the principle of least privilege — our IAM roles request only the minimum permissions necessary.

All data is encrypted in transit using TLS 1.2+. Data at rest is encrypted using AES-256. We support data residency requirements and can store your data in your preferred region.

No. Tailpipe uses cross-account IAM roles (AWS) or equivalent mechanisms (Azure, GCP) for access. Your credentials are never transmitted to or stored by Tailpipe.

We follow SOC 2 security principles and are working towards formal certification. Our open-source scripts are auditable, and we can provide detailed security documentation on request.

Yes. The scripts are open source and fully configurable. You can review, modify, and run them through your own change management process before deployment.

Our scripts include rollback logic and detailed error messages. If a script fails partway through, it will clean up any partially-created resources. You can also contact our support team for help.

GovCloud support is on our roadmap. Contact us to discuss your specific requirements and timeline.

Yes. We provide Terraform modules as an alternative to our shell scripts. These integrate with your existing IaC workflows and support the same configuration options.

Integration updates are handled by re-running the latest version of the setup script. Updates are backwards-compatible and will not disrupt existing data collection.

No. Tailpipe is privacy-first by design. We never capture, log, or transmit prompts or responses. Only metadata is collected: model name, token counts, latency, and status.

We support OpenAI (GPT-4, GPT-3.5), Anthropic/Claude (Claude 4, Claude 3.5), Google Gemini, Mistral, and LangChain (via callback handler for any LangChain-compatible provider).

Tailpipe captures token counts (input, output, cache creation, cache read) and maps them to energy consumption and carbon emissions using the SCI-AI specification. Different models and providers have different carbon intensities.

SCI-AI is the Green Software Foundation's specification for measuring the carbon emissions of AI systems. It defines the metrics needed (tokens, model, compute time) and how to convert them to carbon equivalents. Tailpipe captures all SCI-AI required fields.

Minimal. The Python SDK adds negligible overhead. The Lambda integration adds ~50-200ms. The proxy mode adds 0ms latency (telemetry is captured asynchronously).

Tailpipe uses read-only access to your cloud billing and usage data. We follow the principle of least privilege — our IAM roles only request the minimum permissions needed.

Telemetry data can be stored in your choice of region: EU, US, APAC, or in your own private infrastructure. We support data residency requirements for regulated industries.

We follow SOC 2 security principles and are working towards formal certification. Contact us for our current security documentation.

The baseline assessment (£2,000 per CSP account) covers a comprehensive 3-month analysis including: full carbon footprint calculation, breakdown by service and region, prioritised reduction recommendations, and an executive summary report.

Yes. We offer discounts for organisations with multiple cloud accounts. Contact us for a custom quote.

We offer a pilot programme for qualifying organisations. Contact us to discuss your requirements.