Skip to content

Getting Started

DefendableCloud is the hosted proof vault for AI and agentic work.

DefendableOS is the engine. DefendableCloud is the vault. The engine runs the rulebook; the vault stores and shares the proof.

The central workflow is:

Inputs -> Evidence -> Execution -> Checks -> Verdict -> Approval -> Receipt

A Run is not “the model said it is good.” A Run is evidence plus a rulebook plus a verdict plus an approval plus a hash-chained receipt.

Start here:

Start here:

Start here:

Terminal window
defendable auth login --email [email protected]
defendable auth verify <TOKEN-FROM-EMAIL>
defendable projects create --name "Proof run"
defendable flight-sheets ls --lane agent
defendable runs new --project <project-id> --flight-sheet cre_memo_dscr_ltv_v1
defendable evidence add <run-id> --kind note --label "source facts" --content "..."
defendable submission add <run-id> --output-file ./agent-output.json --agent claude-code --model claude-opus
defendable audit run <run-id>
defendable audit finalize <run-id>
defendable approval set <run-id> --decision approved
defendable receipt generate <run-id>
defendable verify <share-url-or-token>

Current live/foundation surfaces:

  • DefendableCloud marketing site, Vault app, API, and CLI
  • authenticated organizations and roles
  • projects and Runs
  • evidence, submissions, audits, approvals, receipts
  • public receipt sharing and ledger verification
  • dataset library and download receipts
  • model catalog and model pin receipts
  • cook runner routes
  • incidents and stack assessment
  • enterprise runbooks and WAF guidance

Roadmap surfaces:

  • enterprise SSO/SAML/OIDC
  • SCIM
  • external security review
  • formal SOC2 audit
  • public SLA/status program
  • generated OpenAPI reference
  • DDEED public anchoring beyond the current per-org receipt-chain verification

Best sequence:

  1. DefendableCloud Overview
  2. The Defendable Run
  3. CLI Guide
  4. Current API Surface
  5. Enterprise Security