Deep‑local research is a Python agent that answers hyper‑local questions by prioritizing native‑language, local sources over generic English/global results. It discovers local outlets, searches recent local news, and runs a complementary general web search with aligned include/exclude domain filters, iteratively excluding any English sources that slip in (up to four retries). Non‑English excerpts are translated via DeepL; the agent synthesizes concise English summaries with inline citations and streams progress as it works. The system favors parallel tool calls for independent steps (e.g., site discovery plus general search; batched translations), budgets total tool usage to stay under 15 calls, and warns as a soft cap of 20 tool‑call turns approaches—providing partial results and proposing a narrowed follow‑up if the cap is reached. The result is fast, transparent, and locality‑faithful research for place‑specific queries.
Built With
- agents
- deepl
- linkup
- openai
- python
Log in or sign up for Devpost to join the conversation.