Your AI wrote it. Mushi Mushi tells you why it broke.

No. If you build with Lovable or Bolt, you add Mushi Mushi by pasting one prompt, with no terminal. When a bug comes in, you read a short note in plain English and paste the fix instructions back into Lovable or Bolt. You check the change before you publish it.

About five minutes. In Lovable or Bolt, you sign up, copy your script tag and ask the builder to add it. In Cursor or Claude Code, it is one command or one click. A test report gets its plain-English diagnosis in about 10 seconds.

A bug-report inbox with an AI that explains each bug. People who use your app can report a problem with one tap. Mushi Mushi turns each report into a short, plain-English explanation of what went wrong and how to fix it, written as instructions you paste into Lovable or Bolt, or that the AI assistant in your code editor reads itself. It can also check your live site once for free, with no sign-up, for pages that error, dead links and images, and a title still saying "Lovable App". The name comes from 虫 (mushi), the Japanese word for "bug" or "insect".

Yes. Sign up free, copy the script tag your setup page shows (your project ID and key are already in it), and ask Lovable or Bolt to add it to the <head> of index.html. Your published app gets a report button. Each report gets a plain-English diagnosis and a fix prompt you paste back into Lovable or Bolt.

When someone sends a report: what they typed, a picture of the screen if they attach one (they see it first and can remove it; password and card-number boxes are blacked out), the page address, their browser and screen size, and the app's recent error messages. It also counts page visits with a random ID, and skips browsers that ask not to be tracked. No name or email is collected unless your app passes one. On the free plan, reports are deleted after 14 days.

The full list, and one sentence for your own privacy note →

Ask Lovable or Bolt: "Remove the Mushi Mushi script tag from index.html", then publish. The report button is gone and nothing more is sent. To delete what was already collected, delete the project on the Manage apps page of your Mushi Mushi console.

How to remove it, step by step →

No. Mushi Mushi writes the diagnosis with its own AI, and the free plan watches one app with 100 diagnoses a month, no card. Your editor's AI (Cursor, Claude Code, Lovable) does the fix, with the prompt Mushi Mushi hands it.

Built by one developer in Japan (Kensaurus 合同会社), who runs it on his own live apps and sites. The hosted service is live, the libraries are published on npm, and the source code is public on GitHub. The native iOS, Android and Flutter libraries are still previews.

It can be, or it can run alongside Sentry. Sentry is built around what the code threw, with a User Feedback widget and replay alongside. Mushi Mushi starts from what the user reported, ingests Sentry's errors too, explains each one in plain English, and hands your agent a fix prompt to start from. One queue, with or without Sentry.

No. Mushi Mushi works on its own. If you already use Sentry (a popular service that records errors when code crashes), you can point its alerts at Mushi. Those errors then land in the same list as user reports, with duplicates merged, each one explained in plain English and marked resolved in Sentry when the fix is merged.

It runs a setup wizard on your computer. The wizard works out which framework your app uses (React, Next.js, Vue, Nuxt, Svelte, Angular, React Native, Expo, Capacitor or Node), installs the matching Mushi Mushi library, saves your settings in your env file (and warns you if that file is not in .gitignore), and asks before it adds the setup snippet to your code. It never commits, pushes or touches your Git history, and its source is public on GitHub.

Connect your editor to Mushi Mushi with the install box near the top of this page: one command for Claude Code, one click for Cursor. It uses MCP, the standard way AI editors connect to outside tools. Bug reports then arrive as plain-English diagnoses your AI assistant can read, with a fix prompt ready to apply without leaving the editor.

Yes. The libraries you add to your app (SDKs) are MIT-licensed and the server is AGPLv3. You can run the whole thing on your own server with one command, or use the hosted free plan: one app, 100 diagnoses a month, no card.

Yes, if you want it to. Connect your GitHub repo and turn on automatic fixes, and an agent opens a pull request for a diagnosed bug. The built-in fixer needs no key of your own; to use Cursor agents, add your Cursor key, or run them on your own computer with mushi work. With your own Anthropic key, a second AI then reads that pull request next to the original report and says merge, fix first, or do not merge, so a fix that only hides the symptom is caught. Nothing merges unless you merge it or turn on automatic merging.

Web: React, Vue, Svelte, Angular, and any other site through the browser widget. Mobile: React Native and Capacitor. Server: Node. One wizard installs any of them. Native iOS, Android and Flutter SDKs are previews and install from the GitHub repo.