hn · Show HN · 09-30 20:34 47 分 ← 返回列表

Show HN: Perspica – A semantic diff for reviewing code

I started using this a little over half a year ago to help me navigate through some large AI coded PRs people were submitting, because I felt like if I looked at the PR on Github I would just auto pass it through because I didn't want to deal with it. Thought I'd clean it up and share it in case this style of review was useful to anyone else. Obviously it relies on an LLM run to get true semantic groupings of what changes were done, and why, but can also be used without an LLM to get more mechan

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：审查 AI 生成的大量 PR 时，普通 GitHub diff 难以快速理解变更语义和意图，容易草率通过。

目标用户：使用 Claude Code/Codex 等 AI 编程工具的开发者、需要审查 AI 生成 PR 的工程团队和开源维护者。

证据：I felt like if I looked at the PR on Github I would just auto pass it through because I didn't want to deal with it. / help me navigate through some large AI coded PRs people were submitting / reads your prompts and marks which changes you asked for and which the agent decided on its own as well

变现：开源核心 + 团队版订阅，按席位或私有仓库收费；提供 CI 集成、审查审计、私有部署和企业级 Claude Code/Codex 日志解析。

MVP：2 周内做一个 GitHub Action 或 VS Code 插件：解析本地 Claude Code/Codex 会话，对 PR diff 按语义分组，标注用户请求变更与代理自主变更，并生成审查顺序清单。

竞品：CodeRabbit、Greptile、Graphite、Cursor Bugbot、GitHub Copilot code review、Difftastic、Semantic Diff

摘要：Perspica 是一个面向 AI 生成代码 PR 的语义 diff 工具，作者因不想细看大型 AI PR 而开发，强调能读取 Claude Code/Codex 会话并区分用户要求与代理自主改动。HN 发布仅 10 分、4 评论，尚无付费或收入证据，但痛点真实且独立开发者可快速做 MVP。商业化需验证团队是否为 AI 代码审查工具付费。