This paper documents how TwelveLabs integrates with Guidance for a Media Lake on AWS and outlines the operational decisions teams should validate before production deployment, including invocation path, embedding store, prompt strategy, quotas, and index lifecycle management. It covers the pipeline nodes that ship with the Media Lake repository, the invocation paths through Amazon Bedrock and the direct TwelveLabs API, the embedding stores through Amazon OpenSearch and Amazon S3 Vectors, and the operational tradeoffs that matter once real workloads hit the system.