BlackLake for Finance & FinOps
Cap and attribute every dollar of AI spend.
AI cost is unbounded by default. One uncapped agent can exhaust a month’s LLM budget in an afternoon. BlackLake puts cost controls where they belong — at the governance layer, before the spend, attributed to the team that asked.
The problem
AI cost that is attributed after the fact is not a control.
Most cost dashboards aggregate spend after LLM calls have already fired. A denial after the invoice is an observation, not a control. The budget you need to enforce has to act before the model call — not after.
BlackLake’s govern() call evaluates cost at decision time. If an AI Actor would push a budget past its hard limit, govern() returns a denial before the LLM sees the request. Finance gets attribution. Engineering gets the feedback loop. No surprise invoices.
Capabilities
Cost controls built into the governance layer
Per-call cost across seven providers
Cost captured for every governed call across Anthropic, OpenAI, Bedrock, Vertex, Foundry, Gemini, and Ollama. One ledger regardless of which provider each AI Actor uses.
Budgets that deny pre-spend
Set budgets scoped to workspace, AI Actor, tool, or user. When a budget is exhausted, govern() returns a denial before the LLM call fires — not after the invoice arrives.
Soft budget alerts at 50/80/100%
Webhook alerts when a budget crosses your soft threshold. Your finance team sees a spend signal before a hard stop, not as a surprise at end-of-month.
Cost attributed to the AI Actor
Every dollar is attributed to the AI Actor, tool, model, user, and capture path that produced it. Finance reporting that answers 'which team spent what on AI?'
Signed CSV and NDJSON exports
Export the full cost ledger as signed CSV for finance and procurement, or NDJSON for SIEM ingestion. Both are HMAC-signed with the workspace chain.
Cost bound to every receipt
v2 decision tokens cryptographically bind the cost to the receipt. Historical totals stay stable when provider prices change — pricing snapshots are versioned.
Put cost controls on your AI spend today.
Free while in beta. Budgets, cost attribution, and signed exports — all from one workspace with no infrastructure burden.