Deep research agents with enforced budget and local data privacy
Mole is a free open-source terminal deep research agent that caps spending, attaches sources to claims, and keeps local data like CSV on-device. It targets common frustrations: cost overruns, jumbled sources, and unsourced confident answers.
目标用户
Researchers, analysts, and developers who perform deep research from the terminal and need cost control, citation reliability, and privacy for local data.
潜在需求
A way to run deep research with a hard budget limit, every claim traceable to a source, and local data never transmitted to external services.
发生场景
When using AI research agents for deep research, users often face runaway spending, mixed-up sources, and confident but unsourced answers, and they cannot tell whether local data in prompts leaves their machine.
来源证据
AI research agents often overspend budgets, mix up sources, give confident but unverified answers, and raise privacy concerns for local data used in prompts.
Doing research with agents is fun until they blow way past budget, jumble the sources, and don't even give you the best possible answer, just sound confident. And if you want to run some research task on local data - you have no idea where your data ends up after the prompt consumes it. So I built this tool: a deep-research agent with an enforced budget, verified quotes, and a privacy boundary for local data. 1. Never spend more than you budgeted (measured overshoot is 0%). 2. Every claimhttps://news.ycombinator.com/item?id=49303046
为什么值得留意
This signals demand for cost, trust, and privacy guardrails as AI research agents multiply. The tool being free and open source suggests the problem is real but not fully solved by commercial options.
未满足部分
- 主流 AI 研究代理缺少强制预算上限和本地数据不出机器的保证
- 代理常给出无来源的自信答案，且预算容易超支
可能延伸 · 模型推测
- 扩展到其他数据类型（如 JSON、数据库、API 数据）
- 提供团队共享的预算管理和审计日志
- 集成到 CI/CD 或数据处理管道
目前未知
- 该项目刚发布，HN 热度低，实际采用情况未知
- 作者是工具开发者，所述问题可能来自自身经历而非广泛用户反馈
- 预算控制 0% overshoot 的声称未独立验证
继续核实
- 预算超支和来源混乱是否是 AI 研究代理用户的普遍痛点？
- 用户对本地数据隐私的担忧程度如何？是否愿意因此牺牲云端模型能力？
- 强制预算和引用验证是否会让深度研究代理在回答质量上有所妥协？