Recommended Free Tools
The Code Exorcist is a horror-themed AI bug-triage app: it takes an error trace and a relevant code snippet, asks an AI service to propose a diagnosis and fix, then routes that proposal to a person for a decision in Sanity Studio. The AI does not autonomously approve or apply a fix in the workflow Vidisha Gupta describes.
What The Code Exorcist does
Project author Vidisha Gupta describes The Code Exorcist as an early structured step after a bug report arrives. A developer provides an error or stack trace plus the relevant buggy code. An AI service is asked to identify a likely root cause, assign a bug category and threat level, and propose a fix. The result enters a review workflow rather than being treated as an accepted repair.
That is the project’s intended use, not evidence that it improves a team’s speed or that its diagnoses are accurate. Gupta’s account is a description of one implementation, not a benchmark or independent evaluation. Read Gupta’s project account on DEV Community.
How the human-review workflow works
Gupta says each haunting document links to a separate workflowState document. The reported state progression is uncontained → pending_human_review → banished. A history attached to the state record captures an actor, action, timestamp, and notes.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#1 Best Overall
Reviewer decisions in Sanity Studio
Custom Sanity Studio document actions provide two choices: approve the proposed fix with the project’s “BANISH” action, or send the case back for re-analysis. As Gupta puts it: “A human reviewer then makes the final call — directly from Sanity Studio — by clicking either 💀 BANISH (approve the fix, close the case) or 🔄 Send Back for Re-analysis (reject it, reset the case to uncontained).”
These controls make the boundary clear: the AI proposes a diagnosis and patch description; the reviewer decides whether to accept that proposal into the workflow. Human approval is a process decision, not proof that the diagnosis or patch is correct. The account does not establish that the app applies code changes automatically.
Frontend updates
The project account says decisions are logged and reflected on the frontend in real time. Gupta reports that the frontend listens to a GROQ query through Sanity’s client.listen() API, so it can respond to changes in the workflow data.
Technologies Gupta reports using
The described stack includes Next.js, React, TypeScript, Sanity Studio v3, custom document actions, Sanity’s real-time client.listen() API, Groq AI, Tailwind CSS, and Vercel deployment. These are the technologies named in the author’s account; it is not an independently audited inventory of the live app.
Build problems the author encountered
Gupta also describes several troubleshooting episodes. They are useful as project-specific implementation examples, not independently reproduced findings.
- Model names returned errors: the author reports that
llama-3.3-70b-versatileandllama-3.1-8b-instantreturnedmodel_not_found, after which the project switched toopenai/gpt-oss-20b. This records what happened in that build; it is not a guarantee of current provider availability or setup guidance. - Sanity returned a 401: Gupta attributes one authentication issue to the development server not reloading a changed
.env.localfile. - Studio would not start: a duplicate
statusschema field was reported as the cause. - Vercel’s Linux build exposed import problems: case-sensitive imports failed, bringing unused starter files to light.
What this project account can—and cannot—show
The Code Exorcist is a concrete example of an AI-assisted bug intake and review flow: combine a trace with code, request a structured diagnosis, and put the proposal in front of a human reviewer. Gupta positions it as “the moment right after a bug report lands” when a team wants a structured first pass before committing review time. That is product intent, not a measured claim about speed.
Rank #4
The DEV Community account supplies no independent test results or quantified evidence for diagnostic accuracy, time saved, adoption, or reliability. It therefore supports describing the workflow and the author’s build experience, but not claims that this approach reliably finds bugs or is safer or faster than another method.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
