The Console
The surfaces of the ArmorIQ Console at tools.armoriq.ai - Sessions, session detail, tools catalogue, analytics, AI chat, Policy Studio, and Team.
The Console is the read surface at tools.armoriq.ai. It renders the evidence the backend already recorded; it does not enforce or compute decisions itself.
| Surface | What it shows | Slice |
|---|---|---|
| Sessions | Every agent work session, with real token counts and estimated cost | features/sessions |
| Session detail | Tools called in order; allow / deny / hold per call; per-call latency; failures and retries; which rule triggered; an ingest-health indicator | features/sessions |
| Tools catalogue | Every connected tool, filterable by guardrail (ArmorClaude / ArmorCodex), evidence type (configured / observed / guarded), and status; each row shows connection proof (API key, effective policy, whether a real pre-tool evaluation was recorded) | features/mcp |
| Analytics | Real usage and cost, with a guardrail on-vs-off comparison toggle | features/analytics |
| AI Chat | Natural-language Q&A over workspace data, answered with source cards; says "unavailable" rather than inventing an answer | features/chat |
| Policy Studio | Plain-language rule authoring (folder restrictions, tool-level permissions, principal-based policies) | features/policy-studio, features/policies |
| Team | Workspace membership with admin-controlled invites | features/team |
The Console carries no server of its own; it is a static SPA over the backend API, configured with a single env var, VITE_BACKEND_URL.
Architecture
How ArmorIQ Tools is put together - the guardrails on the developer machine, the SDK, the backend enforce and telemetry path, and the read-only Console.
Enforcement
The four decisions - Allow, Deny, Hold, Error - and how the backend enforce path resolves every tool call an assistant attempts.