A small collection of runnable, agent-driven Python projects. Each folder contains a focused workflow with its own entrypoint, data, and logs.

Note: the examples given are based on droidrun version 0.4.0+. To see examples of older version see the legacy/ folder.

• LinkedInJobsScraper

  • Agentic workflow that searches LinkedIn for roles, evaluates matches, and prepares tailored applications.
  • Stores results in jobs/, uses sample configs like candidate_data.json, and records run histories under trajectories/.

• LinkedInLeads

  • End-to-end lead discovery and enrichment for LinkedIn companies and roles.
  • Scrapes company pages, generates role hypotheses, saves outputs in data/companies/, and logs runs in trajectories/.

• TwitterPost

  • Finds trending topics, drafts posts, and generates images to publish on X/Twitter.
  • Includes agents for trend mining, content creation, image generation, and posting, with example assets under images/ and run logs in trajectories/.

• play2048

  • Automated agent that plays the 2048 game on play2048.co, attempting to reach the 2048 tile using strategic reasoning and decision-making.

To learn more about Droidrun, take a look at the following resources:

• Droidrun Documentation - learn about Droidrun features and references
• Benchmark - Our latest benchmark scores.

You can check out the Droidrun GitHub repository - your feedback and contributions are welcome!

We welcome fixes, docs, and new example projects.

• See the full guidelines in contribution.md.
• Keep examples self-contained with main.py, requirements.txt, and an agents/ folder.
• Include a short README in any new example and sample data/assets if needed.
• Include a demo video which describes what you created.