// the find
Oneirocom/Magick
Magick is a cutting-edge toolkit for a new kind of AI builder. Make Magick with us!
Magick is a visual, node-based environment for building LLM agents and data pipelines in TypeScript. You wire nodes into graphs ("spells") and run them against connectors such as Discord, Twitter, and Twilio, with vector search included. It suits people who want chatbots or agents without writing the orchestration code, but the last push was June 2025, so treat it as a snapshot rather than a maintained project.
Spells serialize to plain JSON, so a graph can live in git, be diffed, and be shared as a file. The Wait For All node gives explicit join semantics for parallel branches, which is a sensible answer to the fact that trigger fan-out has no guaranteed order. The embedder app separates ingestion from the agent runtime: it exposes packs, loaders, and a query endpoint behind token auth, and the server migrations store embedding model and provider per agent, so retrieval is part of the data model. Graphs can be embedded as subgraphs and reused, which matters once a team has more than a handful of spells.
The last push was 2025-06-24, more than a year before this review. In an LLM stack where model names, SDKs, and connector APIs change every few months, that gap is a real adoption risk, and nothing in the repo suggests active upkeep. The README sells harder than it documents: installation is three per-OS guides linked out to docs/, and the unbuilt connectors (Zoom, Google Meet, Reddit, Slack) sit in the same feature list as the shipped ones, so check the node library before assuming a capability exists. Running it means an Nx monorepo with about a dozen apps, including two cloud-agent services that look like hosted-product code, on top of Postgres with pgvector. That is a lot of moving parts for a self-hosted tool.