{"platform":{"name":"CAIN TRUST FABRIC","positioning":"AI trust infrastructure for autonomous systems.","core_message":"AI can now act. CAIN Trust Fabric makes those actions provable, governed, and enforceable.","pipeline":[{"actor":"THE AGENT PROPOSES.","detail":"Any runtime -- LangChain, LangGraph, LlamaIndex, CrewAI, an OpenAI/Google/Microsoft/AWS agent stack, or your own loop -- submits an intended action or plan over HTTP or MCP."},{"actor":"CAIN DECIDES.","detail":"The control plane resolves the caller's identity, evaluates policy, scores risk, and asks the verification layer whether the plan can violate the tenant's constraints. One decision record, one verdict."},{"actor":"MCPGATE ENFORCES.","detail":"The decision lands on an enforcement boundary that sits in the call path -- hosted at the CAIN Studio gateway, or self-hosted inside your own network as MCPGate."},{"actor":"THE SYSTEM EXECUTES.","detail":"Authorized calls proceed to the real upstream service. Blocked calls never reach it, and the caller is told which stage stopped them and why."},{"actor":"EVIDENCE REMEMBERS.","detail":"Every decision is written to a durable, tenant-scoped decision record with the stage-by-stage verdicts that produced it, retrievable by id."}],"runtime_interop":{"claim":"Runtime-agnostic by interface, not by integration deal.","detail":"The Fabric is reached over plain HTTPS and over MCP (streamable HTTP). Anything that can make an HTTP request or speak MCP calls the same endpoints with the same headers: LangChain, LangGraph, LlamaIndex, CrewAI, OpenAI, Google, Microsoft and AWS agent runtimes, or a custom loop you wrote yourself. We ship a Python SDK and an MCP server; everything else is one HTTP call. There are no vendor partnerships behind this list and none are implied -- it works because the boundary is a protocol, not a plugin.","runtimes":["LangChain","LangGraph","LlamaIndex","CrewAI","OpenAI Agents","Google (Vertex / ADK)","Microsoft (AutoGen / Azure AI)","AWS (Bedrock Agents)","Custom / in-house runtimes"]},"cross_entitlement":"A subscription on CAIN Studio and a license on MCPGate unlock each other through the cross-platform entitlement bridge, so choosing a deployment model is not a choice about which products you get."},"deployment_models":[{"key":"studio","name":"CAIN Studio","model":"Hosted / Managed CAIN Trust Fabric","role":"The managed control plane. We run it; you point your agents at it.","url":"https://cainstudio.online","boundary":"Hosted enforcement boundary at the CAIN Studio platform gateway.","commercial":"Subscription (Free, Professional and Enterprise tiers), per-call metering on metered services.","who":"Teams that want the Fabric working today without operating it.","control_plane":true},{"key":"mcpgate","name":"MCPGate","model":"Self-Hosted / Customer-Controlled CAIN Trust Fabric","role":"The enforcement boundary you run yourself, inside your own network, where the traffic and the evidence never leave your infrastructure.","url":"https://mcpgate.online","boundary":"Customer-operated enforcement boundary in the customer's own network.","commercial":"One-time perpetual license (bundle or per-tool), self-hosted, no per-call metering by us.","who":"Teams whose data, or whose regulator, will not accept a hosted control plane.","control_plane":false}],"trust_fabric_services":[{"key":"identity","name":"Identity","one_liner":"One namespace for every principal in the loop -- agents, humans, and services.","job":"Know exactly which agent, person, or service is making this call, and prove it.","services":{"studio":["agent-id","cainaccounts","a2a-guard","taloscrypt-keymgmt","taloscrypt-pqc"],"mcpgate":[]},"platform_components":[],"gaps":{"mcpgate":"Self-hosted MCPGate does not ship an identity service today. Callers authenticate with their own IdP or license key; ToolWarden's optional cryptographic attestation covers tool-call integrity, not principal identity. Identity as a Fabric domain is currently hosted-only."}},{"key":"policy","name":"Policy","one_liner":"The rules an action has to satisfy, authored per tenant and evaluated per call.","job":"Write down what is and isn't allowed -- including the regulation you're actually subject to.","services":{"studio":["caingovern","compliance-engine","legal-auditor","quorum-i18n","talosdoc-contract","taloslegal-caselaw"],"mcpgate":[]},"platform_components":[],"gaps":{"mcpgate":"Self-hosted MCPGate has no policy-authoring service; policy is whatever your own OPA/config expresses in front of it. Hosted CAIN Studio is where per-tenant policy upload and evaluation ships today."}},{"key":"authorization","name":"Authorization","one_liner":"Whether this principal may take this action, now, within this budget.","job":"Decide per-call: is this agent allowed to spend this, and does a human need to approve it?","services":{"studio":["cainpay","cainbudget","human-in-loop"],"mcpgate":[]},"platform_components":[],"gaps":{"mcpgate":"Self-hosted MCPGate authorizes by license and by ToolWarden's sandbox rules, not by a per-call budget/approval service. Budget-capped and human-approval authorization is hosted-only today."}},{"key":"risk","name":"Risk","one_liner":"How dangerous this input, output, agent, or fleet looks right now.","job":"Score it before you trust it -- drift, bias, hallucination risk, contradiction, anomaly.","services":{"studio":["sentinel","bias-detector","caindrift","hallucination-firewall","context-health","memory-integrity","contamination-scanner","talosdata-quality","metacognitive-enhancer","trust-state-engine"],"mcpgate":["fairgate","trustledger"]},"platform_components":[],"gaps":{}},{"key":"governance","name":"Governance","one_liner":"Who decided, under what mandate, with what dissent recorded.","job":"Show a regulator or a board that the decision had an owner and a paper trail.","services":{"studio":["blizzard-governance","talosops-governance","guardian-scorecard","lexisguardian"],"mcpgate":["quorum_seal","agentscore"]},"platform_components":[],"gaps":{}},{"key":"verification","name":"Verification","one_liner":"Proof that an output, claim, or artifact is correct before anything trusts it.","job":"Check the answer, the citation, the diff, or the derivation -- deterministically, not with another LLM.","services":{"studio":["rag-verifier","veritasengine","derivation-copilot","taloslogic-proof","talos-verify","vericoding-gate","cainschema","crucible","cainbench","topological-data-analysis","trajectorygate"],"mcpgate":["citegate","driftguard","liftgate","verifygate"]},"platform_components":[],"gaps":{}},{"key":"actionproof","name":"ActionProof","one_liner":"Formal proof that a planned action sequence cannot violate the tenant's constraints -- before it runs.","job":"Prove the plan is safe ahead of execution, or get the counterexample that shows it isn't.","services":{"studio":["actionproof","action-firewall"],"mcpgate":["flowgate"]},"platform_components":[],"gaps":{}},{"key":"security","name":"Security","one_liner":"Keeping attackers, injections, and leaks out of the agent loop.","job":"Stop prompt injection, jailbreaks, secret leakage, and vulnerable code reaching production.","services":{"studio":["adversarial-fuzzer","talos-redteam","pii-redactor","shadow-agent-scanner","taloscode-security","taloscrypt-audit","taloscrypt-vault","taloscrypt-store","taloscrypt-mesh","talosguard-intel","talosguard-siem","talosguard-logiq","secops-patcher"],"mcpgate":["leakguard","probegate","mcp_security_scanner"]},"platform_components":[],"gaps":{}},{"key":"execution-enforcement","name":"Execution Enforcement","one_liner":"The boundary in the call path that actually stops or shapes the call.","job":"Make the decision binding -- not advisory -- at the moment of execution.","services":{"studio":["talos-shield","swarm-orchestrator"],"mcpgate":["tool_warden","meshrouter","mcpiverse"]},"platform_components":["CAIN Studio platform gateway (hosted: entitlement, rate limit, OPA hook, fuzzer hook, Trust Fabric decision path)","MCPGate / ToolWarden (self-hosted: sandboxed tool execution and injection screen)","Open Policy Agent sidecar (policy decision point, currently shadow-mode by default)"],"gaps":{}},{"key":"evidence","name":"Evidence","one_liner":"A durable record of what was decided and what happened, retrievable after the fact.","job":"Answer 'prove this action was authorized' months later, with the original verdicts attached.","services":{"studio":["content-provenance","cainwitness","zkproofvault","multi-agent-failure-tracer","agent-debugger","attestation","decision-intelligence"],"mcpgate":[]},"platform_components":[],"gaps":{"mcpgate":"MCPGate's evidence today is QuorumSeal's signed decision trail and each tool's own audit log, not a unified Fabric decision store. The unified, id-addressable decision record is hosted-only today."}},{"key":"observability","name":"Observability","one_liner":"What the system is doing right now, per service and per fleet.","job":"See health, cost, latency, and behaviour while it's happening -- not in a postmortem.","services":{"studio":["observability","talosops-selfheal","latent-inspector"],"mcpgate":["usageledger","mcpwatch","hubgate"]},"platform_components":[],"gaps":{}}],"governed_workloads":[{"key":"build-ship-code","name":"Build & ship code","one_liner":"Write, review, migrate, and refactor the code itself.","services":{"studio":["talos-coder","talos-evolve","taloscode-review","taloscode-migrate","talosdev-refactor","talosdev-archdiff","visual-ide"],"mcpgate":[]}},{"key":"model-infrastructure","name":"Model & ML infrastructure","one_liner":"Quantize, compress, merge, compile, route, and train.","services":{"studio":["talosai-quantize","model-compressor","weight-alchemist","edge-optimizer","sparselogic-compiler","synthetic-fine-tuner","training-sandbox","omegarouter"],"mcpgate":[]}},{"key":"memory-retrieval","name":"Memory, retrieval & knowledge","one_liner":"Persistent memory, real retrieval, and a shared knowledge store.","services":{"studio":["memorymesh","talos-rag","talos-vectorsearch","knowledge-graph","nexusmind"],"mcpgate":[]}},{"key":"reasoning-orchestration","name":"Reasoning, planning & orchestration","one_liner":"Causal inference, search, physical/scene reasoning, and multi-step execution.","services":{"studio":["causal-network","mcts-engine","stcn","autonomous-task-engine","spatial-synthesizer","echoworkforce","causal-intelligence","counterfactual-intelligence"],"mcpgate":[]}},{"key":"commerce","name":"Commerce & distribution","one_liner":"Billing models and marketplaces for AI products.","services":{"studio":["quorum-oaas","agent-marketplace"],"mcpgate":[]}},{"key":"industry","name":"Industry applications","one_liner":"Domain products built on the same platform: finance, energy, logistics, credit.","services":{"studio":["cainloan","caingrid","cainroute","talosfin-statarb","talosfin-earningsai","arbitrage"],"mcpgate":[]}},{"key":"platform-ops","name":"Platform operations","one_liner":"Running the platform itself: experiment analysis and failure intake.","services":{"studio":["feedback-engine","experimentation-notebook"],"mcpgate":[]}}],"tiers":{"fabric":{"key":"fabric","name":"Trust Fabric Services","blurb":"The trust core. These are the services that decide, enforce, and remember -- the eleven domains the Fabric is made of."},"workload":{"key":"workload","name":"Governed workloads","blurb":"Products that run on the Fabric and are governed by it, but are not part of the trust core: codegen, ML infrastructure, memory, reasoning, and industry applications. Listed separately on purpose -- calling a credit underwriter a trust service would inflate the fabric with something that isn't one."},"site":{"key":"site","name":"Site components","blurb":"Parts of the website itself rather than sellable trust services."}}}