What Ships in AeionClaw — Complete Feature Catalog
A complete inventory of native AI infrastructure. Every persona, skill, reaction trigger, schedule type, memory category, and budget enforcement primitive — enumerated. Use this page to answer "does AeionClaw do X?" with certainty.
What Powers AeionClaw
AI Engine
Streaming SSE responses, multi-model orchestration across 57 providers, persona auto-routing by conversation context, tool dispatching to 89 MCP tools, conversation lifecycle management.
Intelligence Orchestration
Knowledge retrieval via semantic embedding, context assembly per turn, conversation summarization at checkpoint, fact extraction, persistent memory across sessions.
Reaction Engine
Event-driven automations — subscribe to any platform event, filter via the full Aeion QueryBuilder operator set, dispatch any registered MCP tool, cooldown tracking, per-day quotas.
Workflow Executor
Multi-step automation chains with variable interpolation, error handling, conditional branching, dry-run mode for safe iteration.
Site Monitor
Proactive monitoring — page health, uptime tracking, anomaly detection, alert dispatching.
Schedule Processor
Cron-driven recurring tasks — 5 task types × 4 frequencies × 4 delivery channels.
Queue Backbone
BullMQ workflow job queue with concurrency control, retry strategy, dead-letter handling.
8 Built-In AI Personas — Auto-Selected by Context
AeionClaw automatically picks the right persona based on conversation context. Each ships with a domain-specific system prompt and scoped toolset, so an order-status question never wastes tokens loading CRM context, and vice versa.
47 AeionClaw Skills — One Per Module
Every module ships its own AeionClaw skill — a packaged, callable expert that knows the module's workflows and best practices. See `/platform/claw/skills-catalog` for the full categorized matrix.
Event-Driven Reactions — Automation Without Workflows
The Reaction Engine monitors your platform's activity and fires AI actions on matching events. Any platform event × the full Aeion QueryBuilder operator set × the platform MCP tool registry = ~2,400 possible automations out of the box.
Scheduled Tasks — Recurring AI Operations
Run intelligence on a cron. Daily morning briefings, weekly executive reports, monthly KB sweeps — all defined declaratively.
Smart Memory — Cross-Session Context
AeionClaw doesn't forget. Every conversation feeds a fact-extraction pipeline that builds a per-user memory store, retrievable via semantic embedding similarity.
Execution Budgets — AI Spend Governance
Set hard caps on AI spend per tenant. Warn / soft-block / hard-block enforcement based on percentage thresholds.
Complete Control & Governance
Role-Based Scope Limiting
AeionClaw respects the user's platform role. A "Viewer" cannot trick the AI into executing "Delete."
AI Spend Budgets
Strict token + cost budgets per tenant (Daily / Weekly / Monthly caps).
Skill Factory
Users draft custom Skills, test in sandbox, submit for admin approval before global deployment.
Immutable Action Logs
Every AI tool call recorded in append-only audit log with full inputs
Learn more