Question 1

What is nable?

Accepted Answer

nable builds cost controls into the AI you already use. It is an MCP server that makes Claude, Cursor, and your AI agents cost-aware: ask about your AWS, Azure, GCP, and AI or LLM spend, and your agents check changes against your budget before they act. It runs on your own machine and your credentials never leave it.

Question 2

Is nable free?

Accepted Answer

Yes. The Dev tier is free with no credit card and no expiry: cost queries, anomaly detection, rightsizing, LLM spend tracking, and every connector. Paid tiers add remediation pull requests, alerts, scheduled digests, with managed single-tenant hosting available as an optional add-on, contact us for a demo.

Question 3

Does nable see or store my cloud credentials?

Accepted Answer

No. nable runs on your machine. Credentials stay in your OS keyring and cost data caches in a local SQLite database. There is no nable backend that holds your data, and nothing is shipped to a vendor.

Question 4

Can nable change my cloud infrastructure on its own?

Accepted Answer

No. nable is propose-only. It drafts a pull request or opens a ticket with the fix, and a human reviews and applies it. It never edits, deletes, or buys anything in your environment autonomously.

Question 5

What clouds and tools does nable support?

Accepted Answer

AWS, Azure, GCP, and Kubernetes, plus more than ten SaaS and AI providers including Datadog, Snowflake, Databricks, Stripe, OpenAI, Anthropic, and Amazon Bedrock. It exposes 180+ tools your editor can call. None of them can change your infrastructure: fixes ship as pull requests or tickets a human approves.

Question 6

How is nable different from Vantage, CloudHealth, or the AWS FinOps agent?

Accepted Answer

nable is local-first, your credentials and bills never leave your machine; AI-native, it lives in Claude or Cursor instead of a separate dashboard; and genuinely cross-cloud, including AI and LLM spend in the same answer. It proposes fixes as pull requests for human approval rather than acting on its own.

Question 7

What is a FinOps MCP server?

Accepted Answer

MCP, the Model Context Protocol, lets AI editors call external tools. A FinOps MCP server exposes cloud-cost tools to your AI editor, so you ask about spend in your own words and the editor calls the right tool. nable is a local-first FinOps MCP server.

Question 8

Does nable normalize cost data across providers?

Accepted Answer

Yes. Every provider is normalized into FOCUS 1.2, the FinOps Foundation's open billing standard. nable extends the standard past AWS, Azure, and GCP to usage-based SaaS like Snowflake, Datadog, MongoDB Atlas, and Databricks, and to per-model AI spend from OpenAI, Anthropic, OpenRouter, and LiteLLM with token counts preserved. One query spans your whole bill because every provider answers in the same schema.

Question 9

Can nable show what my AI coding costs?

Accepted Answer

Yes. nable attributes merged pull requests and commits to the AI model that wrote them and joins your LLM spend by model, so you can see what each model shipped and what it cost per pull request or per commit.

Your cloud and AI bill,answered in your editor.

Your cloud and AI bill,
answered in your editor.