Glossary · Cost engine

Marginal cost per run

The cost of one more execution of a workflow. On most browser infra it’s flat; on Twin it falls as cache hits replace LLM calls.

What is marginal cost per run?

Marginal cost per run is what it costs to execute a workflow one additional time. On infrastructure that re-runs the LLM every execution, that cost is roughly constant no matter how often you run. On Twin, repeated and re-phrased runs increasingly resolve through the semantic cache to deterministic replay, so the marginal cost trends toward browser time alone — toward zero LLM spend.

Why it matters

It’s the number that decides whether automating at volume is sustainable. A flat marginal cost means your bill scales linearly with usage; a falling one means cost per 1,000 runs drops as you scale — the entire reason Twin exists.

See it in context: read how Twin compiles and replays a run, follow the cost-cutting guide, or browse use cases and comparisons.

Run your first skill

Give an LLM agent a real browser, compile the workflow once, and watch the marginal cost fall as the cache takes over.