Skip to content

perplexity-at-home

perplexity-at-home packages a set of LangGraph research workflows behind a clean Python package and CLI. The repo currently includes a durable deep-research path, a tighter pro-search loop, and a lightweight quick-search builder for simpler question-answer flows.

What the package gives you

  • Pydantic Settings for OpenAI, Tavily, LangSmith, and Postgres configuration
  • a packaged CLI entrypoint: perplexity-at-home
  • a packaged Streamlit launcher: perplexity-at-home-dashboard
  • optional Postgres-backed LangGraph persistence for checkpoints and store state
  • packaged builders for quick-search, pro-search, and deep-research graphs
  • a dashboard surface for switching between all three workflows
  • runnable demos under examples/
  • unit and integration coverage around settings, graph construction, persistence, the CLI surface, plus gated live E2E coverage for real external services

Core commands

make help
pdm run perplexity-at-home --help
make deep QUESTION="What is Tavily?"
make up
make db-setup
make deep-persistent QUESTION="What is Tavily?"
make dashboard
make test-e2e

Documentation map

  • Getting Started: local setup, environment, and run flow
  • Architecture: graph layout, persistence, and package boundaries
  • Reference: API and runtime entrypoints pulled from the package