Applied AI / Insights
A useful control needs a known boundary. Start by identifying which requests travel through it.
01 / Map
List the agent, provider, credentials and request route. A limit inside a gateway cannot enforce a budget on calls sent directly to the provider. Subscription usage and metered API spend also need separate treatment.
02 / Bound
Decide what the limit protects and what the user sees when it is reached. Test that behaviour with controlled inputs before relying on it for a long run. Treat pre-run cost ranges as estimates, not promises.
03 / Verify
Runcap documents routed spend controls and a separate Proof Gate for coding workflows. Its captured token-reduction example is one measured call, not a universal saving.
Open source, demonstration PRs and evidence04 / Decide
A spend ceiling does not establish whether the work is correct. Define the acceptance evidence, protected checks and human-review conditions separately.
Explore AI reliability & evaluationA practical first step
Discuss the bottleneck, the tools you already use, and what a worthwhile result would look like. No system access needed to start.
Review my workflowApply the guide
Continue the investigation
Applied AI
AI agents now write a growing share of the code that ships. That shift moves the hard question from "can the agent write the change?" to "can I trust what it did, and did it cost what I expected?" Those are two different failure modes, and…
Read the analysis ↗Applied AI
Almost everyone budgets AI the same wrong way. They look at the price of one call - a few cents, maybe a fraction of one - multiply by how many times they expect to run it, and conclude the whole thing is cheap. Then the monthly bill…
Read the analysis ↗