Agents pick providers by what's wired in, not by what the call costs. Provider routing fixes that: every search, extract, or browse call walks a ladder — free and local first, cheap paid next, premium only when nothing cheaper returns a result.
These posts cover the mechanics: how routers decide, how fallback chains behave when a rung returns zero hits, and why the default provider for any capability should cost $0. Start with the complete guide to provider routing.
Outages are the risk everyone plans for and the least likely one to hurt you. Quality drift, repricing, and quota cuts arrive without status pages.
The free-cheap-premium ladder isn't a search trick. Extraction, transcription, embeddings, and code execution all have the same shape — here's how to build the ladder.
None of these show up as errors. They show up as an invoice that's 3–6× what the same traffic should cost, one reasonable-looking call at a time.
The same ladder idea, two different physics. Why a router that treats search results and model completions identically gets both of them wrong.
Cost was the wedge. Routing is the product. What's shipping, what changed, and what Frugal Cloud will add.
One captured run, walked rung by rung: 0 hits in 529ms, a fall-through, 3 hits at $0, and exactly what lands in the ledger afterward.
Paid providers should be the fallback, not the default. The case for making your agent's first call free, with the failure modes stated honestly.
A backup provider you've never exercised is a config value, not a plan. The rules I use to order rungs, avoid paying twice, and break flapping providers out of the chain.
The term gets used for two different things. Here's what a router actually contains — policy, health, fallback, accounting — and the honest cases where you should skip one.
Agents pick providers by what got wired in, not by price. Here's how a routing layer walks every tool call down a ladder — free first, paid as fallback.