Build TypeScript AI agents with memory, tools, workflows, and RAG.
npx skills add https://github.com/hashintel/hash --skill mastra
Please help me install this Agent Skill. Skill: mastra Source: https://github.com/hashintel/hash/tree/main/.codex/skills/mastra Command: npx skills add https://github.com/hashintel/hash --skill mastra
Mastra provides a TypeScript framework for building AI agents with memory, tools, workflows, and retrieval-augmented generation (RAG).
Install Mastra and follow the Getting Started guide at mastra.ai/docs/v1/getting-started/installation.
High-intent search queries and answers about installing and using this skill.
Build AI agents using Mastra, a TypeScript framework providing memory management for conversation history, deterministic tool orchestration, and workflow coordination. Install Mastra packages and define structured tools with inputSchema, outputSchema, and execute functions to create memory-enabled agents.
Mastra supports memory adapters including LibSQL, MongoDB, Postgres, and Upstash, plus vector stores like pgvector, Pinecone, Qdrant, and MongoDB. Choose based on your infrastructure to enable persistent agent memory and semantic search across your RAG workflows.
Implement RAG by processing and chunking documents, configuring vector embeddings, and performing cost-aware vector search through Mastra's RAG capabilities. Mastra handles document ingestion, embedding integration, and vector store coordination end-to-end.
Yes, Mastra supports deterministic multi-step workflows using .then() and .commit() methods. Chain tool calls, streaming, and generation modes together to create complex agent reasoning pipelines with explicit control flow.
Mastra includes MCP client and server integration, enabling your agents to communicate with external tools and services through the Model Context Protocol. This extends tool availability and interoperability beyond native Mastra tools.
Streaming mode outputs results progressively as they're generated, while generation mode waits for complete results before returning. Select based on latency requirements and whether real-time feedback or final outputs matter for your agent workflow.
MassimilianoPiliConfigures LangChain4J vector stores for RAG applications with multiple databases.
efoo-teamQuery Langfuse REST API traces, observations, scores, and datasets.
neuron-coreIntegrate vector stores, embeddings, and document loaders into Neuron AI agents.
BermudaLocalsStore embeddings and metadata locally for fast similarity search.
