Settings/Agent Engine & AI Intelligence

Autonomous Agent Calibration & LLM Configuration

Gemini 2.5 Native

Configure default foundational reasoning models, provide encrypted Gemini API credentials, and toggle autonomous dispatch vs. fail-closed gates.

Foundational LLM Architecture

Google Cloud Vertex / AI Studio

Gemini 2.5 FlashRecommended

Ultra-low latency sub-second inference with 768-dim embeddings. Optimized for reactive inbound triage and high-throughput email replies.

Active Selection

Gemini 2.5 ProDeep Reasoning

Complex enterprise contract reasoning, multi-turn nuance extraction, and high-precision knowledge synthesis.

Gemini 1.5 FlashLegacy Fast

High volume cost-efficient baseline model for basic lead classification and auto-tagging.

Custom Gemini API Key

Encrypted at rest with AES-256-GCM

If left blank, NousHub will use the platform default managed Gemini credentials. Supply your Google AI Studio key (`AIza...`) to route tokens through your personal billing quota.

Managed Quota

Execution & Safety Gate Calibration

Deterministic Guardrails

Draft-Only Gate (Fail-Closed)

Investor Pitch Mode

All synthesized replies are held in the Daily Review Queue with held_by_safety_lock: true. Requires human operator review and dispatch via Hotkey A.

Risk Level: Zero Autonomous Hallucination Risk

Autonomous Auto-Dispatch

Sub-3m SLA

Inbound leads scoring below 80 priority points are automatically synthesized and dispatched immediately via the Gmail API. High-priority enterprise leads (≥ 80) remain held for executive review.

Routing: Priority-Aware Triage Matrix

Stylistic Persona & Tone Calibration

Prompt Conditioning Vector

Agent Persona Name

Professional Title / Signature

Formal (Executive Triage)30% CasualCasual (Warm & Conversational)

Concise (Sub-75 words)25% DetailedDetailed (Comprehensive In-Depth)

Core Brand Voice Philosophy