NewEvery agent tool call now routes through the governed MCP gateway — policy check, audit log, then dispatch.Read the changelogarrow_forward
XOOZLYAI Orchestrator
routeHow it workssmart_toyAgent fleetshieldGovernance & trustsellPricingmenu_bookDocsarticleBloghistoryChangelog
Log inStart free
grid_viewProductsrouteHow it workssmart_toyAgent fleetshieldGovernance & trustsellPricingmenu_bookDocsarticleBloghistoryChangelogHelp centreLog in
Start free
  1. XOOZLY
  2. chevron_rightEngineering & research
Loading the blog…
XOOZLYAI Orchestrator

Ready to orchestrate your AI fleet with total confidence?

Start a project, set a cost cap, and watch the loop run with every tool call policy-checked and logged before it leaves the platform.

Start freearrow_forwardSee what shipped
  • check_circleNo credit card required
  • check_circleYour keys, your model spend
  • check_circleEvery tool call audited

Platform

  • All productsarrow_forward
  • Orchestration looparrow_forward
  • Agent fleetarrow_forward
  • Governed control planearrow_forward
  • Pricingarrow_forward

Trust & legal

  • Trust centrearrow_forward
  • Security & compliancearrow_forward
  • AI use policyarrow_forward
  • Report a vulnerabilityarrow_forward
  • Data subject requestsarrow_forward

Resources

  • Documentation & guidesarrow_forward
  • Engineering & research blogarrow_forward
  • Changelogarrow_forward
  • Uptime & reliabilityarrow_forward

Browse by domain

  • Strategy & Researcharrow_forward
  • Developmentarrow_forward
  • QAarrow_forward
  • Governance & Orchestrationarrow_forward
  • All 10 domainsarrow_forward
XOOZLY
© 2026 XOOZLY · 61 agents · 10 domains · 8 SDLC stationsAI use policyarrow_forwardLog inarrow_forward99.9% uptime SLAarrow_forward
  1. XOOZLY
  2. chevron_rightEngineering & research
  3. chevron_rightDispatch
Loading this dispatch…
  1. XOOZLY
  2. chevron_rightEngineering & research
  3. chevron_rightEliminating runaway inference costs with hard project-level caps
FinOps

Eliminating runaway inference costs with hard project-level caps

A cap that is reconciled after the fact is a report. A cap that is read before the billable call is a boundary. The difference shows up the first night an agent loops.

FOFinOps EngineeringverifiedCost caps & span accounting·Jul 19, 2026·5 min read
𝕏in
On this page
  1. 01Checked before, not reconciled after
  2. 02Span-level accounting is what makes the check cheap
At a glance
Worked example — one project's monthAuto-pause threshold 80%
Spent to date · 78%Hard ceiling · 100%
Architectural questions?

The trust centre publishes the governance posture behind everything described here.

Open the trust centrearrow_forward
01

Checked before, not reconciled after

A project's cap is a hard limit read before a run starts, not a number compared against an invoice at the end of the month. A run that would start past the cap does not start, and a run that crosses it while executing is paused rather than allowed to finish.

End-of-month reconciliation is a reporting feature. It cannot stop a recursive loop at three in the morning, because by the time the report exists the money is spent.

02

Span-level accounting is what makes the check cheap

Cost is attributed per span as a run proceeds, so the balance a cap is checked against is already current. There is no aggregation pass to wait for, which is what makes a pre-flight check viable on a hot path.

FOWritten by the FinOps EngineeringCost caps & span accounting. Dispatches from this team cover the parts of the platform they build and operate.

Related in the product

  • arrow_forwardPlans & cost controls
  • arrow_forwardThe governed control plane
Governed by default

Ready to govern your autonomous fleet?

Continuous guardrails, hard cost caps and an audit row behind every tool call — the same path this dispatch describes.

Start freearrow_forwardRead the trust centre
Suggested reading

More from the engineering blog

Explore all dispatchesarrow_forward
  • GovernanceAug 10, 2026

    Why we built a central MCP gateway instead of trusting agents directly

    Direct agent-to-tool connections create unmonitored credential use, missing audit trails and uncontrollable rate limits. Here is how one centralized proxy enforces policy, writes the audit row and dispatches — in that order — before a request reaches a third-party server.

    verified_userONE PATH PER TOOL CALLPolicy check → audit log → connector dispatchNo domain may bypass it
    8 min readRead article: Why we built a central MCP gateway instead of trusting agents directlyarrow_forward
  • AI OpsAug 03, 2026

    Inside the Rule Optimizer: proposals a human still has to approve

    How detected failure patterns in autonomous runs become proposed rule edits a human approves — and why approval writes a new version instead of editing one in place.

    Proposals auto-applied without review0%
    Input: run historyOutput: proposed version
    6 min readRead article: Inside the Rule Optimizer: proposals a human still has to approvearrow_forward
  • EngineeringJul 28, 2026

    Deterministic guardrails vs. probabilistic LLM evaluators

    Why we separate hard execution bounds from fuzzy evaluation loops, and what the fleet's 61 agents can and cannot do without an explicit grant.

    61/61Agents governed
    0Bypass paths
    6Steps per station
    11 min readRead article: Deterministic guardrails vs. probabilistic LLM evaluatorsarrow_forward