Agent skill · product strategy · jeremylongshore
klingai-reference-architecture
Production reference architecture for Kling AI video generation platforms.
Why this skill is useful
Provides detailed architecture patterns and scaling strategies specific to Kling AI video generation that the AI wouldn't reliably generate on its own.
What it needs
About 4k tokens when loaded. Last updated 2026-08-07. 2,604 stars on the source repository.
What this skill does
Kling AI Reference Architecture Overview Production architecture for video generation platforms built on Kling AI. Covers API gateway, job queue, worker pool, storage, and monitoring layers. Architecture Diagram Component Details API Layer Worker Service Scaling Guidelines Component Scaling Strategy ----------- ----------------- Workers Scale by queue depth (1 worker per 3 concurrent API tasks) API servers Horizontal, behind load balancer Redis Single instance for <1K jobs/day, cluster for more Storage S3/GCS scales automatically CDN CloudFront/Cloud CDN for global delivery Concurrency Limits by Tier Tier Max Concurrent Tasks Workers Needed ------ --------------------- ---------------- Free 1 1 Standard 3 1 Pro 5 2 Enterprise 10+ 3-4 Docker Compose Setup Prerequisites Defined availability, latency, retention, residency, and cost objectives; a threat model; and named owners for policy, data rights, operations, and publication approval. A secret manager, private staging storage, immutable artifact digests, queue-level idempotency, and a bounded model/credit/concurrency allowlist. Synthetic or rights-cleared fixtures for load and integration tests. Production likeness or customer media requires consent and an explicit processing purpose; test runs must not export contacts or source media. Instructions 1. Keep the API gateway responsible for authentication, authorization, prompt and provenance validation, content-policy checks, destination allowlists, and budget estimation before queueing work. 2. Put only opaque job references and approved parameters on the queue. Workers obtain short-lived credentials from the secret manager, enforce idempotency, and submit a private draft rather than publishing directly. 3. Start each release with a watermarked sandbox canary. Verify policy, source rights, suppression/destination rules, output integrity, aggregate error rate, quota, and cost before an owner approves staged promotion. 4. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills jeremylongshore/klingai-reference-architecture--96f9d1