Original Summary

Hey all - I made a thing with my clanker Kevin: Ototo (<a href="https:&#x2F;&#x2F;ototo.dev&#x2F;" rel="nofollow">https:&#x2F;&#x2F;ototo.dev&#x2F;</a>) - <i>yes this is clanker driven dev</i>. The idea is to offload code exploration and other token&#x27;y heavy tasks to a smaller model to do the work and chuck the result back. This helps in a couple of ways (1) it saves tokens if you can offload to a cheaper or free local model (2) it can keep the main context clearer as it removes all the sherlocking thinking&#x2F;exploration. It also saves more tokens than using anthropics own exploration agent.<p>I had originally played around with a graph based approach but found this to work better. Currently, I have this running with Qwen 3.8 27B on my framework (I had Kevin build me a custom inference engine to squeeze all the juice I could out of the AMD - <a href="https:&#x2F;&#x2F;gitlab.com&#x2F;handmadedigital&#x2F;projects&#x2F;shocho" rel="nofollow">https:&#x2F;&#x2F;gitlab.com&#x2F;handmadedigital&#x2F;projects&#x2F;shocho</a>).<p>It&#x27;s been through a bunch of tests (results are available if people want) against a load of different repos, including comparing subsequent commits of large repos with claude, claude + exploration agent and claude + ototo and so far ototo seems to really help.<p>The site is very &#x27;LLM&#x27;y at the moment - I haven&#x27;t had time to edit it to make the text more readable for us humans. Sorry.<p>Hope it helps folks in their day-to-day.<p>--EOL

中文概览

中文标题： Show HN：Ototo——把代码探索外包给更小模型的工具

作者做了 Ototo，思路是把代码探索等耗 token 的任务交给较小模型完成再把结果送回，以节省 token 并让主上下文更干净，据称比 Anthropic 自带探索代理更省 token。目前在本机用 Qwen 3.8 27B 和自建推理引擎运行，并在多个代码库上做过测试。

  • 情报分类：技术学习与提效
  • 分类依据：面向开发者的代码探索与省token效率工具
  • 信息来源：Hacker News 新项目
  • 发布时间：2026/10/6 05:25:51