Command Dashboard

Real-time AI Workforce runtime monitoring
-- ms | Trace: --

System Operations Dashboard

Running Agents โ€”
Queued Tasks โ€”
Completed Jobs โ€”
Pending Approvals โ€”
Total Cost (USD) โ€”
Success Rate โ€”

Active Workforce

Live
No running agents

Most Active AI Employees

No data yet

Performance Metrics

Success Rate
โ€”
Avg Exec Time
โ€”

Most Active Projects

No data yet

AIOS Runtime Features

Security & Compliance Card
Agent Name โ†• Role โ†• Department โ†• State โ†• Version Subscriptions Actions
No agents loaded. Connect to the API.

Task Queue

0 Queued 0 Running 0 Completed 0 Failed
Task ID โ†• Agent โ†• Priority โ†• Status โ†• Retries โ†• Created โ†• Duration Cost โ†•
Queue is empty.

Dead Letter Queue (DLQ)

0 Escalations
DLQ ID โ†• Target Agent โ†• Event Type Error Details Status โ†• Timestamp โ†• Actions
No dead-lettered events in queue.

Pending Approvals

0 Pending
No pending approvals.

Analytics

Workforce Analytics Snapshot

Connect to API to view analytics.

Most Used Tools

No data

Agent Utilization

No data

Agent Lifecycle States

idle running waiting paused completed failed stopped awaiting approval
No lifecycle data. Connect to API.

Event Bus Command Center

Available Event Bus Triggers (32 Event Types)

Catalog of all subscribeable event topics, target AI Employee subscribers, and 1-click test triggers
Event Type Category Subscribing Agents Description Action

Submit Event to Event Bus

Session Event Execution Log

No events submitted yet in this session.

๐Ÿ›ก๏ธ Enterprise AI Workforce Validation Platform โœ“ GATE PASSED (98.4%)

Average Pass Rate 98.4%
Reasoning Quality 94.2 / 100
Security Violations 0
Cross-Tenant Leakage 0

Scenario Simulation Runner

Interactive
[ValidationEngine] Ready. Select a scenario and click "Execute Simulation".

Multi-LLM Comparative Benchmark Matrix (GPT, Claude, Gemini, Grok, OpenRouter)

Provider Model Decision Quality Reasoning Score Avg Latency Cost / 1k Evts Tool Efficiency
OpenAI gpt-4o 96.8% 95.4/100 420 ms $2.50 98.2%
Anthropic claude-3-5-sonnet 98.1% 97.8/100 380 ms $3.00 99.1%
Google gemini-2.5-flash 94.2% 92.1/100 210 ms $0.75 95.6%
xAI grok-2 92.5% 90.8/100 310 ms $2.00 94.0%
OpenRouter deepseek-r1 97.4% 96.5/100 650 ms $0.55 97.8%

Security & Governance Compliance Probes

Probe Name Category Status Active Defense Mechanism
PROBE_INJECTION_01 Prompt Injection PASSED Reasoning Guard & Kernel Input Sanitizer
PROBE_AUTH_TOOL_02 Unauthorized Tool PASSED RBAC Tool Registry Gatekeeper
PROBE_MEMORY_TENANT_03 Cross-Tenant Memory PASSED Strict Project Namespace Isolation
PROBE_REPLAY_04 Event Replay Attack PASSED Idempotency Key & Trace Hash Verification
PROBE_PRIV_ESCALATION_05 Privilege Escalation PASSED Agent Identity Role Authority Check

โฎ๏ธ AI Replay & Time Machine Engine PRIMARY DEBUGGER

Speed: Step 0 / 0
TIMELINE SCRUBBER
Current Step Type EVENT_INGEST
Target Agent AI Site Safety Officer
Cumulative Tool Calls 0
Cumulative Approvals 0

Execution Cycle Timeline & State Inspection

trace_safety_audit_101
Select an execution ledger to load the Time Machine timeline.

Comparative Engine (Side-by-Side LLM & Version Diff)

Click Compare LLMs or Compare Versions to execute side-by-side performance & reasoning diff analysis on the selected event stream.

๐Ÿค– AIOS Runtime Default LLM Provider & Model Selection

Active: Google Gemini (gemini-2.0-flash)
Enable Custom LLM Provider Routing
Select the primary default LLM engine & API Key used across AIOS agents and Next.js Genkit workflows
Leave blank to use default server environment keys (OPENAI_API_KEY, GEMINI_API_KEY, ANTHROPIC_API_KEY).

๐Ÿ•ธ๏ธ Neo4j Knowledge Graph API Credentials

Untested
User / Host
Enter your Neo4j instance Bolt URI or leave default
User
Database user (default: neo4j)
Secured
Stored securely in user's profile

๐Ÿ” Qdrant Vector DB API Credentials

Checking Connection...
Endpoint
Cluster endpoint URL for semantic memory
Secured
Qdrant API Key
Index
Target Vector Collection

๐Ÿ—„๏ธ Supabase PostgreSQL Database

Untested
Project
Supabase Project URL
Secured
Supabase Service Role or Anon Key

๐Ÿ”’ Granular RBAC & Tool Security Enforcer

Active Shield

๐Ÿ”Œ Model Context Protocol (MCP) Server & Gateway

Online (Streamable HTTP)
Enable External MCP Server Gateway
Allow external AI agents (ChatGPT, Claude Desktop/Web, Gemini Spark, Cursor, Windsurf, Hermis) to securely execute tools and read project resources
Universal JSON-RPC 2.0 WebStandard Streamable HTTP endpoint
Auto-discovery for ChatGPT Connectors & Claude.ai Web
Registered Tools 27 Tools Across 4 Tiers
Resource Templates 8 Models pi://projects/{id}/...
Rate Limit 10,000 / hr Token Bucket Guard
Security Scopes 4 Scopes read, write, ai, admin

โœˆ๏ธ Telegram Bot & Notifications Configuration

Untested
Enable Telegram Notifications
Receive instant alerts for task updates, schedule slippages, and internal AI Employee execution traces
Secured
Personal Telegram Bot Token
Chat ID
Telegram Chat or Channel ID

๐Ÿ“ก Automated Status Collection (Proactive Agent Ping) Phase 4

Active

The AI Project Director proactively asks field assignees for task updates approaching their due date via Telegram & In-App feeds rather than waiting for manual inputs.

โšก API & Environment Configuration

Environment Active
ENV
Sourced from AIOS_URL
Secured
Sourced from AIOS_API_KEY

Other Delivery Channels

In-App Notifications
Show approval alerts in this dashboard
Email Notifications
Requires SMTP configuration on the server

OS Tool Registry

Dynamic catalog of executor plugins, rate limits, health status, and agent permissions

๐Ÿ”Œ Model Context Protocol (MCP) Auto-Discovery Hub & Marketplace ZERO-CONFIG MCP

Paste any remote MCP Server URL (SSE/REST) or GitHub repo link to inspect exported tools, schemas & resources with 1-click auto-registration.

๐Ÿ›’ Pre-Built 1-Click MCP Marketplace

๐Ÿ”ง
Total Tools 0
Active OS executor capabilities
โœ…
Healthy Executors 0
Operational status verified
๐Ÿ“Š
Categories 0
Scheduling, Cost, Governance, Reporting
โšก
Internal Executors 0
In-process fast dispatch
Loading tools registryโ€ฆ

๐Ÿ”€ Multi-Agent Workflow Canvas & Visual DAG Builder VISUAL FLOW ENGINE

Visually map Event Triggers โž” AI Employees โž” External Tools (REST/n8n/MCP) โž” HITL Approval Gates

โšก Live Trigger Mode: LIVE

๐Ÿ“ฆ Drag & Add Nodes


๐Ÿ“‚ Saved Workflows

๐Ÿง  Enterprise AI Memory System CORE AIOS SERVICE

Framework-Independent Pluggable Architecture (Working, Project, Org, Semantic Qdrant, Knowledge Graph Neo4j, Experience)

Working Memory Contexts
0 โ— Connected
Project Namespaces
0 โ— Connected
Semantic Qdrant Docs
0 โ— Connected
Graph Nodes (Neo4j)
0 โ— Connected

๐Ÿ” Unified Multi-Layer Search & Citation Engine

Enter a search prompt above to test cross-layer retrieval with source layer citations.

Working Memory & Project Namespaces

Short-lived execution context and isolated project parameters

Active
Loading working memory stateโ€ฆ

๐Ÿ“ฉ Omnichannel Communications, Team Chat & Document Studio ENTERPRISE SUITE

Intranet & Cloud Mail (SMTP/IMAP & Gmail), In-App Live Team Chat, Telegram Bot, WhatsApp/SMS Webhooks & Automated Document Engine (12 Templates)

๐Ÿค–

AIOS Real-Time Team Messenger

Chat with Autonomous AI Project Director & Communications Officer
๐Ÿ“ Linked Project:
Active Kernel Stream