Mastra
Mastra is a TypeScript framework for AI applications and agents. It brings agents, graph-based workflows, human-in-the-loop, memory, evaluation and observability together in one stack for Node.js environments.
- Project
- Mastra
- Publisher
- Mastra
- Repository
- github.com/mastra-ai/mastra
- Documentation
- mastra.ai/docs
- Languages
- TypeScript
- Category
- AI agent frameworks
What is Mastra?
Mastra is a framework for building AI applications and agents in TypeScript and embedding them in React, Next.js or Node.js applications or running them as a standalone server. The core is released under the Apache 2.0 licence; code in directories named ee/ is covered by an enterprise licence that requires a valid licence for production use.
Core capabilities
- Model routing
- One standard interface to more than 40 model providers, including OpenAI, Anthropic and Gemini.
- Agents
- Agents choose tools, iterate and stop when the model returns an answer or a stopping condition is met.
- Workflows
- A graph-based workflow engine with .then(), .branch() and .parallel() for explicitly controlled multi-step processes.
- Human-in-the-loop
- Agents and workflows can be suspended until input or approval is given; state is stored and resumed later.
- Context, evals and observability
- Conversation history, retrieval from your own sources and memory; built-in evals and observability; agents can be exposed as MCP servers.
When Mastra fits
Mastra fits when your team works in TypeScript and agents are to become part of an existing web or Node.js application. The Mastra Studio development environment is used to build, test and manage agents, workflows and tools.
Using Mastra effectively in an AI-first business model
Mastra offers two paths: open-ended agents that decide on tools themselves, and workflows with a fixed sequence. Which path suits a task follows from its risk – the greater the impact of an action, the stronger the case for defined steps and an approval point.
The built-in evals only pay off with criteria derived from the business goal. Define what a good run looks like and check it regularly in operation, not only before launch.
From tool to outcome
In AI-first coaching you clarify which outcomes AI agents should be accountable for in your go-to-market and within which framework. REVAGO then designs, builds and runs your system. You pay for results – not for activity.
Frequently asked questions
What is Mastra?
Mastra is a TypeScript framework for AI applications and agents. It includes model routing, agents, graph-based workflows, human-in-the-loop, memory, evals and observability and can be embedded in React, Next.js or Node.js applications.
Is Mastra open source?
The core and the vast majority of the code are released under the Apache 2.0 licence. Code in ee/ directories is source-available under the Mastra Enterprise License; it may be used freely for development and testing, but production use requires a valid licence.
How does Mastra support human-in-the-loop?
Agents and workflows can pause and wait for input or approval. Mastra stores the execution state so that the process can resume at the same point, even after a long time.
Further reading in the wiki
A neutral overview, not a recommendation of individual providers. REVAGO is not affiliated with the listed projects. Descriptions are based on the projects’ public information; features and licences may change.