by hanxiao · Agent Tool · ★ 42
Searchbox Given a query, a dataroom, and a token budget. The local in a minimal Pi harness explores the dataroom with local tools such as bash, grep, embeddings, rerankers in an airgapped loop, then answers. Live demo → hanxiao.io/searchbox Why Everyone who knows me knows I'm super test-time-com"pilled". In my view, search is test-time compute (TTC): you wire trained embeddings, rerankers, multi-vector retrievers, query expanders into a pipeline at test-time to squeeze out relevancy. Don't scale TTC, say a keyword search hands you the answer, and it's probably not good enough.
| Stars | 42 |
| Forks | 5 |
| Language | Python |
| Category | Agent Tool |
| License | MIT |
| Quality Score | 65.3604902571638/100 |
| Last Updated | 2026-06-26 |
| Created | 2026-06-09 |
| Platforms | python |
| Est. Tokens | ~16k |
These tools work well together with searchbox for enhanced workflows:
Explore other popular agent tool tools:
searchbox is Airgapped closed-corpus QA loop: a self-hosted Qwen3.6 agent explores a .zip dataroom under a token budget with local tools. It is categorized as a Agent Tool with 42 GitHub stars.
searchbox is primarily written in Python. It covers topics such as agentic-search, apple-silicon, closed-corpus-qa.
You can find installation instructions and usage details in the searchbox GitHub repository at github.com/hanxiao/searchbox. The project has 42 stars and 5 forks, indicating an active community.
searchbox is released under the MIT license, making it free to use and modify according to the license terms.