Mastra
TypeScript framework for building AI-powered applications with Mastra
Core Concepts
Agents
Create AI agents with tools, memory, and multi-agent coordination
Browser
Browser automation with AgentBrowser and Stagehand
Editor
Build and manage prompts with the Mastra Editor
Evals
Evaluate agent performance with built-in and custom scorers
MCP
Model Context Protocol for extending agent capabilities through a standardized protocol
Memory
Memory gives your agent coherence across interactions and allows it to improve over time
Models
Configure and use LLM providers including OpenAI, Anthropic, Google, and Mistral
Observability
Monitor and debug your Mastra application with logging, tracing, and metrics
RAG
Retrieval Augmented Generation for grounding AI responses in external knowledge
Server
Deploy Mastra as a standalone server with various adapters and authentication
Streaming
Stream responses from agents and workflows for real-time interactions
Voice
Voice agents with text-to-speech and speech-to-text capabilities
Workflows
Workflows provide structured ways to define complex AI-powered processes
Workspaces
Workspaces provide file system and sandbox capabilities for agents