Skip to content

Example Gallery

Each example is executable from the repository and covered by tests. No example depends on a paid service unless explicitly marked “real model mode.”

Native DeerFlow-style workbench

Location: examples/deerflow/

Capabilities:

  • selective task execution
  • computed selectors
  • scope isolation
  • streaming through the real Driver
  • long-term store
  • checkpoints
  • time travel / restore
  • HITL resume
  • vector search
  • DOT export
  • browser E2E

Run offline tests:

uv run --directory examples/deerflow --extra test pytest -q

Open the local workbench:

uv run --directory examples/deerflow python frontend/server.py --port 8000

Optional real-model mode uses an AGNES-compatible OpenAI endpoint and is skipped when AGNES_API_KEY is absent.

Tutorials

The docs tutorials are runnable scripts, not pseudocode. CI checks each script and every performance or behavior claim points back to a reproducible command:

uv run --directory python/reactivegraph python scripts/check-tutorials.py

Start with:

  1. docs/tutorials/01-why-reactive.md
  2. docs/tutorials/02-first-graph.md
  3. docs/tutorials/03-computed-scope.md
  4. docs/tutorials/04-streams-resume.md
  5. docs/tutorials/05-durability.md
  6. docs/tutorials/06-1000-nodes.md

Real DeerFlow base-replacement PoC

The real bytedance/deer-flow factory path was compared against a ReactiveGraph replacement:

uv run --directory python/reactivechain   python -m pytest $PWD/benchmarks/deerflow_real/ -q

See benchmarks/deerflow_real/REPORT.md.

Migration example

The LangGraph importer test contains a complete upstream graph → native graph example:

uv run --directory python/reactivechain \
  python -m pytest tests/test_langgraph_import.py -q

See migration guide for the API.

Production example

Generate a production-ready Dockerfile, health check, and Compose lifecycle:

pip install \
  "git+https://github.com/from000/reactive-graph.git#subdirectory=python/reactivegraph" \
  "git+https://github.com/from000/reactive-graph.git#subdirectory=python/reactivegraph_cli"
reactivegraph new my-app
cd my-app
reactivegraph dockerfile .
reactivegraph up .

See production deployment guide.