Research Scientist stack

For ML researchers and computational scientists. LlamaIndex for ingesting paper PDFs, Modal for serverless GPU runs, Weaviate for hybrid keyword+vector retrieval, Langfuse for tracking experiment provenance, MCP fetch so the agent can pull data from public APIs mid-session. The fetch + memory MCPs together turn Claude Code into a meaningful long-running research assistant.

What's in this stack

• Claude Code
  Anthropic's terminal-native coding agent with MCP-native tool use and skills system.
  npm_package
• Anthropic
  Claude model family — the default Claude Code provider. Sonnet, Opus, Haiku.
  env_key
• LlamaIndex
  Data framework for LLM apps — strongest in document ingestion and RAG pipelines.
  pip_package
• Fetch MCP Server
  Official MCP reference server for HTTP fetching and HTML-to-markdown conversion.
  mcp_pip
• Memory MCP Server
  Official MCP reference server providing a persistent knowledge-graph memory across sessions.
  mcp_npm
• Modal
  Serverless GPU and CPU for AI workloads — Python-native, zero infra setup.
  pip_package
• Weaviate
  OSS vector DB with hybrid BM25+vector search and modular embedding adapters.
  pip_package
• Langfuse
  OSS, self-hostable LLM observability — traces, evals, prompts, and datasets.
  pip_package

How to use it

Click any entity above for its install instructions. Most are one-line installs; the MCP servers are JSON-config additions to your Claude Code config file.

The free Arcade tier lists these entities for browsing + per-entity install. Arcade Pro adds 1-click bulk-install: with one button in the desktop app, the entire stack lands configured and ready.

← Browse all arcade categories · data/arcade.json