Infrastructure
Loading…
Authenticated session
Not connected
Control Plane usability
Limited
Tasks awaiting you
–
Loading…
Active missions
–
Loading…
Agents eligible
–
Loading…
InfrastructureLoading…
Mission executionNot loaded
Eligible runtimesNot loaded
Providers verifiedNot loaded
Owner actionsNot loaded
Last verifiedNot loaded
Tokyo AI Core
…
Status
Version
–
Environment
–
Plugins
–
Features
–
Last check
–
Tokyo World
Checking…
CORE
–
PLAN
–
FORGE
–
AIRLOCK
–
ARCHIVE
–
Read-only visualization of real Tokyo agent/mission state.
Provider Status
loading…
Services
loading…
Integrations
loading…
Active Missions
loading…
Tasks Awaiting You
loading…
Runtime Fleet
loading…
Tokyo AI
מרחב עבודה פרטי · הקשר פרויקט וזיכרון תחומים
ממתין לחיבור
0 / 8000
התשובות נשענות על מודל מקומי כברירת מחדל. פעולות רגישות עדיין דורשות הרשאה ואישור.
Full Status
loading…
loading…
Autonomous Missions
Live projection of the authoritative mission store (missions.json / worker_runs.json). Statuses are evidence-backed — no estimated progress.
loading…
↳ EXECUTION RESOURCES BACKING THE MISSIONS ABOVE
Runtime Fleet
Mission-execution eligibility, quota, and load — genuinely live, tracked per mission run. Capability facts (installed/enabled) are a static catalog, not a live probe; see Agent / CLI Runtimes (AI Providers tab) for a live auth check of the same runtimes.
loading…
Tokyo World
Read-only visualization of real Tokyo agent/mission/runtime state. No connection = dark. No state = no markers. Optional 3D view available via the toggle below.
loading…
SummaryEvidence-backed surfaces
StatusEvidence-backed views
Important problemOpen a surface to inspect
Action: choose a product area below. Details stay inside each surface.
Owner-facing governance and audit surfaces, in one place.
GOVERNANCE
AUDIT
DIAGNOSTICS
Voice / wake diagnostics
Wake word: Tokyo · unavailable · trained Tokyo wake model is not configured. Voice Mode remains available by explicit microphone action.
Owner action: configure OWW_MODEL with the trained Tokyo wake-model file before enabling hands-free wakeup.
Projects → Tokyo
Read-only composition of existing operational sources. Unknown and untracked evidence remains visible.
loading…
Devices
loading…
Live Workflows
loading…
Token Operations
loading…
Budget
loading…
Audit Log
loading…
Approval Lifecycle
loading…
Feature Flags
loading…
System Readiness
loading…
Model / API Providers
Inference providers, with configuration, authentication and health kept distinct. A configured credential is not proof of credits or readiness.
loading…
Local Models (Ollama)
Real Ollama inventory: installation, load state and Tokyo qualification are shown separately. Rejected and deferred models remain visible.
loading…
Agent / CLI Runtimes
Agent execution is separate from model inference. Live checks cover the Tokyo service process; host-only CLIs remain NOT_VERIFIED until a governed host discovery channel exists. For mission-scheduling eligibility of the same runtimes, see Runtime Fleet (Missions tab) — a separate, quota/load-based authority.
loading…
Phase 21.5 — Checkpoint 7
loading…
Memory System
loading…
Knowledge Graph
loading…
Evaluation Summary
loading…
Recent Evaluations
loading…
Agent Runtime Status
Runtime facts are shown only when verified; architecture definitions are not activity.
loading…