--- url: https://chaoslabs.xyz/cortex/foundry title: "Cortex Foundry: Put your AI work to work" description: "Foundry sends every request down its best-value path, brings proven context to each prompt, and turns repeated work into models you own." crawled_at: 2026-10-17T16:24:15Z ask: https://chaoslabs-xyz.ora.run/ask docs: https://chaoslabs-xyz.ora.run/docs lang: en --- Cortex [Foundry Intervenes and optimizes](https://chaoslabs.xyz/cortex/products) # Choose the intervention that fits the work. Atlas evidence identifies the Work Stream, constraint, success measure, and safety boundary. Foundry then intervenes and optimizes by choosing the right intervention: routing, just-in-time enrichment, distillation against private evals, or semantic security. Foundry intervention result · Support escalation triage · Before Candidate paths are not yet matched to this work. Intervention Route eligible models against policy, quality, cost, and latency. Verified outcome Cortex support v3 is selected after the same support eval. Intervention and Optimization with Foundry ## Use approved work to improve the next request. Route work, bring approved context forward, build evals from reviewed outcomes, and promote a model only after it passes your tests. Cortex · Foundry Live product view ### 01 · Route Which path fits this work? Routing decision · Support escalation triage · Route selected · Work Unit requirements · Policy Approved US-hosted path Quality floor 90 on support eval Cost ceiling $0.01 per task Latency target Under 2 seconds Candidate paths Same task and eval Cortex support v3 91 $0.006 1.7s General model A 84 $0.028 2.3s Frontier model 82 $0.041 3.8s Quality · 91 · Cost · $0.006 · Latency · 1.7s · Route Each Work Unit gets a route selected against policy, quality, and cost. View product Just-in-time enrichment ## Bring the active Work Unit the context it needs. When a Work Unit needs context, Foundry retrieves relevant, permission-scoped evidence at run time. Just-in-time enrichment carries the constraints for the active Work Stream without treating every prompt the same. - Context selected for the active Work Stream - Scoped by team, project, and permission - Available through configured application, API, or MCP paths Context & Memory Semantic routing ## Choose the model path that fits the constraint. For a diagnosed constraint, Foundry uses semantic routing across eligible models, tools, agents, and caches. Agentic governance and semantic security keep the intervention within its approved safety boundary on configured supported paths. - Routing measured against the Work Stream success measure - Agentic governance matches autonomy to the approved boundary - Semantic security applies controls on configured supported paths Routing & Yield Evals and Benchmarks ## Use private outcomes to define the quality bar. Private evals and benchmarks turn accepted, rejected, and verified Work Units into a success measure for a recurring Work Stream. Foundry uses that measure to decide whether a proposed optimization is worth adopting. - Private evals grounded in verified outcomes - Success measures by Work Stream - Public and private benchmarks in one quality workflow Training example: “I was charged twice for one order.” The response says, “I’ve refunded the duplicate charge.” A reviewer correction creates two criteria: cite the policy and state the manager approval requirement. Separate held-out prompt, not used for training: “The delivery was late. Can I get a refund?” Both answers are checked against the same two criteria. Current answer: “Under Refund policy §4, I’ve issued your refund.” The policy is cited, but the manager approval requirement is omitted. Candidate answer: “Under Refund policy §4, I’ll request manager approval before issuing a refund.” The policy is cited and the approval requirement is stated. This checks answer text only, not whether a refund or manager approval occurred. Example answers Model distillation ## Distill repeated work against private evals. When a recurring Work Stream has private outcome and verifier data, Foundry can distill the pattern into a task-specific model. Teams can assess it against the Work Stream's private evals and safety boundary before promotion. - Distillation based on private evals and verifier data - Models deployed in your configured environment - Changes measured against the Work Stream success measure Model Tuning Accepted Work Units → Corpus Step 1 · 12.4M turns • Auto-labeled · SFT + preference pairs · Step 2 Edits vs. Rejects as signal Cortex-triage-v3 · Yours · Owned · Hosted in your VPC ## Choose the intervention that moves the outcome. Turn observed work into a tested change, then verify the result before you promote it. Cortex by Chaos Labs Keep the intelligence your company creates. Atlas ### Atlas Workforce Visibility · Work & Attribution · Foundry ### Foundry Routing & Enrichment · Model Tuning · Security ### Security AI Firewall · Oracle Sentinel · Solutions ### Solutions [CISO / Security & Governance](https://chaoslabs.xyz/cortex/solution-ciso) CFO / FinOps · [CTO / VP R&D](https://chaoslabs.xyz/cortex/solution-cto) · CEO [People / CHRO / L&D](https://chaoslabs.xyz/cortex/solution-people) Platform ### Platform Architecture · Semantic Model · [Docs](https://docs.chaoslabs.xyz/cortex) · Company ### Company Chaos Labs · Testimonials · Legal © 2026 Chaos Labs. All rights reserved.