Category: AI Systems
Category
A category is a broad topic/domain. These pages share category: ai-systems.
Agent Frameworks
general
- Agent Frameworks - AI Systems collection
Browse ZajLibrary pages for AI Systems / Agent Frameworks.
Mastra
- Mastra collection
The TypeScript agent framework — agents, tools, workflows, memory, RAG, eval, and observability, built on top of the Vercel AI SDK. Start here for the mental model, then go deep.
- Concepts handbook
The Mastra vocabulary — every core primitive defined crisply, so the rest of the handbook reads cleanly.
- Architecture handbook
How a Mastra app fits together — the package boundary, the central runtime, and how one request flows from route to model and back.
- Building agents handbook
The minimal end-to-end path — define a tool, define an agent, register it on the Mastra instance, and call it from your app.
- Tools handbook
How an agent acts on the world — typed, schema-validated functions the model can call, with predictable inputs and outputs.
- Workflows handbook
Deterministic, durable orchestration around the model — steps, sequencing, branching, parallelism, and suspend/resume for human-in-the-loop.
- Memory handbook
How agents keep continuity — conversation history, semantic recall, and working memory, all backed by storage and a vector index.
- RAG handbook
Retrieval-augmented generation in Mastra — chunk, embed, store, and retrieve your own documents so an agent answers from your data, not just its training set.
- Multi-agent systems handbook
When one agent isn't enough — a supervisor agent that delegates to specialist sub-agents, plus the other coordination patterns.
- Evals & scorers handbook
Turn "looks fine" into a number — automated quality checks on agent output you can track over time and gate releases on.
- Observability handbook
See what your agents actually do in production — every model call, tool call, and workflow step traced via OpenTelemetry and exported to your tooling.
- MCP handbook
How Mastra interoperates with the wider agent ecosystem — consuming external tools as a Model Context Protocol client, and exposing your own as a server.
- Deployment handbook
From local dev to production — the Mastra CLI, deployers, serverless targets, and what to watch in production.
integrations
AI Agent Integration
- Add a chatbot to your app guide
How ZajLibrary's /chat assistant was built — assistant-ui + the Vercel AI SDK, grounded in the library via RAG, with chat persistence on Postgres.
- Build an MCP server for your docs site guide
How ZajLibrary exposes itself to AI agents over the Model Context Protocol — a build-time content index plus one on-demand Vercel function, with keyword and semantic search.
- Create a Claude Code plugin guide
Package an MCP server, a skill, and a slash command into one installable Claude Code plugin — using the ZajLibrary plugin as the worked example.
- AI Agent Integration - Integrations collection
Browse Learn, Build, and Resources pages for AI Agent Integration.
general
- Integrations - AI Systems collection
Browse ZajLibrary pages for AI Systems / Integrations.
uncategorized
general
- AI Systems collection
Browse ZajLibrary pages in AI Systems.