Mastra

aid: mastra-ai
name: Mastra
description: Mastra is a TypeScript framework for building AI-powered applications and agents, built by the team behind Gatsby
  (Sam Bhagwat, Abhi Aiyer, Shane Thomas). The framework provides production-grade primitives for agents, workflows, RAG,
  tools, memory, evals, voice, and observability, and integrates with 40+ model providers through the Vercel AI SDK. The framework
  is Apache 2.0 licensed open source, with a hosted commercial offering (Mastra Cloud) for deployment, tracing, and team collaboration,
  and an enterprise self-hosted edition for regulated environments.
kind: contract
accessModel:
  pricing: unknown
  onboarding: unknown
  trial: false
  try_now: false
  public: false
  label: Unknown
  confidence: low
  source: []
  generated: '2026-07-22'
  method: derived
position: Producing
access: Open Source
image: https://kinlane-images.s3.amazonaws.com/shared/apis-json/icons/mastra-ai.png
tags:
- Agents
- Artificial Intelligence
- Workflows
- RAG
- MCP
- Memory
- Evaluation
- LLM Observability
- TypeScript
- Voice
- Open Source
created: '2026-05-25'
modified: '2026-05-25'
url: https://raw.githubusercontent.com/api-evangelist/mastra-ai/refs/heads/main/apis.yml
specificationVersion: '0.19'
x-type: opensource
x-tier: 2
apis:
- aid: mastra-ai:mastra-framework
  name: Mastra Framework
  description: The Mastra framework is a TypeScript-first agentic stack distributed as the @mastra/* family of npm packages.
    The core package wires agents, workflows, memory, RAG, tools, MCP, evals, voice, and observability into a single Mastra
    instance that can be embedded in Next.js, Express, Hono, SvelteKit, Astro, Nest, Fastify, Koa, or stood up as a standalone
    server. Models are reached through the Vercel AI SDK so any of 40+ providers (OpenAI, Anthropic, Google, Bedrock, Groq,
    etc.) plug in through one interface. Apache 2.0 licensed.
  humanURL: https://mastra.ai/docs
  tags:
  - Agents
  - Artificial Intelligence
  - Workflows
  - TypeScript
  - Framework
  properties:
  - type: Documentation
    url: https://mastra.ai/docs
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra
  - type: PackageManager
    url: https://www.npmjs.com/package/@mastra/core
  - type: License
    url: https://github.com/mastra-ai/mastra/blob/main/LICENSE.md
  - type: GettingStarted
    url: https://mastra.ai/docs/getting-started/installation
  - type: Examples
    url: https://github.com/mastra-ai/mastra/tree/main/examples
  - type: Templates
    url: https://mastra.ai/templates
- aid: mastra-ai:mastra-agents
  name: Mastra Agents
  description: Mastra agents define behavior with a model, instructions, tools, memory, and processors, then expose .generate()
    and .stream() for typed text, object, and structured output generation. Agents support multi-step tool calls, dynamic
    instructions, runtime context, scoped memory, and Vercel AI SDK v5 message format.
  humanURL: https://mastra.ai/docs/agents/overview
  tags:
  - Agents
  - Artificial Intelligence
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/agents/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/packages/core/src/agent
- aid: mastra-ai:mastra-workflows
  name: Mastra Workflows
  description: Workflows are durable, graph-based pipelines composed from typed steps with branching, parallelism, loops,
    suspension, human-in-the-loop, and event triggers. Step inputs and outputs are validated against Zod schemas and each
    run is persisted so workflows can be replayed, resumed, and observed.
  humanURL: https://mastra.ai/docs/workflows/overview
  tags:
  - Workflows
  - Orchestration
  - Durable Execution
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/workflows/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/workflows
- aid: mastra-ai:mastra-memory
  name: Mastra Memory
  description: The memory package provides conversation history, working memory, and semantic recall for agents, pluggable
    across PostgreSQL, LibSQL, Redis, Upstash, MongoDB, MS SQL, DynamoDB, Cloudflare D1, ClickHouse, Couchbase, and other
    backends. Vector recall integrates with PgVector, Pinecone, Chroma, Qdrant, Astra, Elasticsearch, OpenSearch, Turbopuffer,
    Vectorize, LanceDB, DuckDB, S3Vectors, and more.
  humanURL: https://mastra.ai/docs/memory/overview
  tags:
  - Memory
  - Vector Database
  - Persistence
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/memory/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/packages/memory
  - type: PackageManager
    url: https://www.npmjs.com/package/@mastra/memory
- aid: mastra-ai:mastra-rag
  name: Mastra RAG
  description: The RAG package provides document loaders, chunkers, embedders (including FastEmbed local embeddings), rerankers,
    and retrieval helpers that compose with any Mastra vector store. Supports semantic, keyword, and hybrid retrieval, metadata
    filtering, and graph RAG patterns.
  humanURL: https://mastra.ai/docs/rag/overview
  tags:
  - RAG
  - Retrieval
  - Embeddings
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/rag/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/packages/rag
  - type: PackageManager
    url: https://www.npmjs.com/package/@mastra/rag
- aid: mastra-ai:mastra-mcp
  name: Mastra MCP
  description: First-class Model Context Protocol support — Mastra ships an MCP client for connecting to any MCP server and
    exposing its tools to agents, plus an MCP server implementation for publishing your own agents, workflows, and tools as
    MCP-callable capabilities. Includes a docs MCP server and a registry-of-registries.
  humanURL: https://mastra.ai/docs/mcp/overview
  tags:
  - MCP
  - Model Context Protocol
  - Tools
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/mcp/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/packages/mcp
  - type: PackageManager
    url: https://www.npmjs.com/package/@mastra/mcp
- aid: mastra-ai:mastra-evals
  name: Mastra Evals
  description: Built-in evaluation library with model-graded (LLM-as-judge), rule-based, and statistical metrics for measuring
    agent output quality, hallucination, faithfulness, relevance, bias, toxicity, and answer correctness. Evals run locally
    or against Mastra Cloud and emit traces alongside production runs.
  humanURL: https://mastra.ai/docs/evals/overview
  tags:
  - Evaluation
  - Testing
  - LLM-as-Judge
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/evals/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/packages/evals
  - type: PackageManager
    url: https://www.npmjs.com/package/@mastra/evals
- aid: mastra-ai:mastra-voice
  name: Mastra Voice
  description: Voice abstractions for speech-to-text, text-to-speech, and realtime voice agents with provider adapters for
    OpenAI (including Realtime API), ElevenLabs, Deepgram, Google, Google Gemini Live, Azure, AWS Nova Sonic, Gladia, Inworld,
    Murf, PlayAI, Sarvam, Speechify, ModelsLab, and xAI Realtime.
  humanURL: https://mastra.ai/docs/voice/overview
  tags:
  - Voice
  - Speech
  - Realtime
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/voice/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/voice
- aid: mastra-ai:mastra-observability
  name: Mastra Observability
  description: OpenTelemetry-native tracing for agents, workflows, tools, and LLM calls with first-party exporters for Mastra
    Cloud, Langfuse, LangSmith, Braintrust, Arize, Arthur, Laminar, Datadog, Sentry, ClickHouse, and PostHog. AI SpanKind
    semantic conventions and an OTel bridge are supported out of the box.
  humanURL: https://mastra.ai/docs/observability/overview
  tags:
  - Observability
  - Tracing
  - OpenTelemetry
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/observability/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/observability
- aid: mastra-ai:mastra-cli
  name: Mastra CLI
  description: The mastra CLI scaffolds projects (`npm create mastra`), runs the local dev server with hot reload, opens the
    Studio playground, generates types, runs migrations on memory stores, deploys to Mastra Cloud, and runs evals. Installable
    as @mastra/cli.
  humanURL: https://mastra.ai/docs/cli/overview
  tags:
  - CLI
  - Developer Tools
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/cli/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/packages/cli
  - type: PackageManager
    url: https://www.npmjs.com/package/mastra
- aid: mastra-ai:mastra-client-sdks
  name: Mastra Client SDKs
  description: Client-side SDKs that talk to a Mastra server from the browser, React applications, or any JavaScript runtime.
    Includes @mastra/client-js (universal JS client), @mastra/react (React hooks for agents, streaming, and workflows), and
    @mastra/ai-sdk (adapters for the Vercel AI SDK UI primitives).
  humanURL: https://mastra.ai/docs/client-sdks/overview
  tags:
  - SDK
  - Client
  - React
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/client-sdks/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/client-sdks
  - type: PackageManager
    url: https://www.npmjs.com/package/@mastra/client-js
- aid: mastra-ai:mastra-deployers
  name: Mastra Deployers
  description: Deployer packages that bundle a Mastra app for a target runtime. Built-in deployers cover Mastra Cloud, Cloudflare
    Workers, Vercel, and Netlify. Custom deployers can be written for any Node-compatible platform.
  humanURL: https://mastra.ai/docs/deployment/overview
  tags:
  - Deployment
  - Serverless
  - Edge
  properties:
  - type: Documentation
    url: https://mastra.ai/docs/deployment/overview
  - type: SourceCode
    url: https://github.com/mastra-ai/mastra/tree/main/deployers
common:
- type: DomainSecurity
  url: security/mastra-ai-domain-security.yml
- type: Website
  url: https://mastra.ai
- type: Documentation
  url: https://mastra.ai/docs
- type: GitHub
  url: https://github.com/mastra-ai/mastra
- type: GitHubOrg
  url: https://github.com/mastra-ai
- type: Blog
  url: https://mastra.ai/blog
- type: ChangeLog
  url: https://github.com/mastra-ai/mastra/releases
  name: GitHub Releases
- type: Pricing
  url: https://mastra.ai/pricing
- type: PackageManager
  url: https://www.npmjs.com/org/mastra
  name: npm @mastra Org
- type: Examples
  url: https://github.com/mastra-ai/mastra/tree/main/examples
- type: Templates
  url: https://mastra.ai/templates
- type: Community
  url: https://discord.gg/BTYqqHKUrf
  name: Discord
- type: XTwitter
  url: https://x.com/mastra_ai
- type: LinkedIn
  url: https://www.linkedin.com/company/mastra-ai
- type: YouTube
  url: https://www.youtube.com/@mastra-ai
- type: Course
  url: https://mastra.ai/course
  name: Mastra MCP Course
- type: Book
  url: https://mastra.ai/book
  name: Principles of Building AI Agents
- type: Podcast
  url: https://mastra.ai/agent-hour
  name: Agent Hour
- type: License
  url: https://github.com/mastra-ai/mastra/blob/main/LICENSE.md
  name: Apache 2.0 + Mastra Enterprise License
- type: YCombinator
  url: https://www.ycombinator.com/companies/mastra
- type: Crunchbase
  url: https://www.crunchbase.com/organization/mastra-44ba
overlays:
- type: APIs.io Search
  url: overlays/apis-io-search.yml
- type: API Evangelist Ratings
  url: overlays/api-evangelist-ratings.yml
maintainers:
- FN: Kin Lane
  email: kin@apievangelist.com