Meltano Cloud

Meltano Cloud is a managed platform for moving data from various sources into a warehouse on a schedule. It is built on Meltano and the Singer ecosystem, with 460+ sources and 40+ destinations.

The connector exposes the workflow rather than the API, so the platform can be operated entirely in conversation: someone who has never opened the UI can go from "I want my Stripe data in a warehouse" to a scheduled, running pipeline without leaving Claude.

Key features:

Build a pipeline end to end. Provision an isolated workspace with its own managed warehouse and git repository, search the catalog, read a connector's real settings before configuring it, and create a scheduled pipeline. Installing a plugin is implicit in configuring it.

Run and troubleshoot. Run history, row-count metrics and logs, plus a diagnosis flow that assembles a failed run's context - exit code, configuration, error-log tail - into a grounded prompt, so Claude explains the real failure and proposes a specific fix.

Answer from documentation. A vendored Meltano knowledge base is searchable through the connector, so questions about replication methods, plugin syntax or settings are grounded in docs rather than recollection.

Reliability at a glance. Opt-in health verdicts and a written summary naming the unstable pipelines, so "how is my data platform doing?" is one question.

Security by inheritance. OAuth 2.1 against our existing identity provider, with RFC 9728 protected-resource discovery. Every tool call runs as the signed-in user and inherits the same per-workspace authorization as our REST API, so the connector grants no privilege the user does not already have. Secrets are redacted from every response, logs included, and credentials, SSH keys and repository tokens are never returned.

Config-as-code, both ways. Changes made through the connector are written back to the workspace's git repository, and users who prefer git can edit locally and ask Claude to publish.