Usage Guide¶

This section covers all core capabilities of Echo Agent, from model integration to task scheduling, helping you unlock the full potential of the Agent runtime.

Contents¶

Section	Description
Model Integration	Multi-model configuration, load balancing, custom endpoints
Tools & Permissions	Built-in tool registration, permission policies, sandbox isolation
Execution Backends	Local/remote/containerized execution environments
Browser & Media	Web interaction, screenshots, file upload & media processing
Memory Management	Short/long-term memory, vector storage, retrieval strategies
Knowledge Base	Document import, index building, RAG-enhanced retrieval
Sessions	Parallel sessions, context isolation, session persistence
Tasks & Planning	Task decomposition, execution plans, multi-step reasoning
Scheduled Jobs	Cron expressions, scheduled triggers, failure retries
Dashboard	Web dashboard, real-time monitoring, operation auditing
Cost Control	Token budgets, usage alerts, model fallback strategies

Overview¶

Echo Agent combines the reasoning capabilities of large language models with an orchestrable tool system, delivering a complete Agent runtime. The sections here are organized by functional domain:

Model Layer: Configure one or more LLM backends with support for OpenAI, Anthropic, local models, and other providers.
Tool Layer: Declare which tools the Agent can invoke, and control scope and rate through permission policies.
Execution Layer: Choose where code and commands run — local processes, Docker containers, or remote sandboxes.
Extended Capabilities: Browser automation, media processing, and knowledge base retrieval give the Agent richer perception and action.
State Management: The memory system and session management ensure context coherence across multi-turn, multi-session scenarios.
Scheduling & Monitoring: Scheduled jobs, the Dashboard, and cost controls help you run Agents safely and efficiently in production.