🤓 Latest Submissions
HealthBridge — A Clearer Path to Care
Patients often arrive at appointments with scattered reports, concerns, and unanswered questions. HealthBridge is a healthcare handoff prototype that helps patients and clinicians prepare for a more informed conversation. Patients enter concerns in their own words and upload documents. HealthBridge extracts text from supported files, organizes records, and generates a structured visit brief with links back to its source evidence. Conflicting or uncertain information stays visible instead of being silently resolved. An assigned clinician can inspect sources, correct records with an audit trail, approve or flag a brief, and publish plain-language or reviewed translated instructions. Patients can acknowledge instructions and export a portable summary. If underlying information changes, previously generated briefs are marked outdated. Built with React, TypeScript, FastAPI, and MongoDB (with SQLite for local development), the prototype supports English, Hindi, and Kannada interfaces. Its optional model-backed evidence selector validates selected excerpts against source documents; the default source-only mode does not run an LLM. HealthBridge uses synthetic demo data only. It does not diagnose, prescribe, or replace professional judgment. Its goal is a clearer, traceable, human-reviewed patient-to-clinician handoff.
18 Oct 2026
Doctor Dev
A repository can pass its existing tests and still be fragile. Important error and authorization paths may have no test evidence, while environment variables, Docker settings, CI, runtime versions, ports, and README instructions can disagree. Developers often discover these gaps only during onboarding, review, or deployment because each tool checks a different piece of the system. Doctor Dev gives developers one place to inspect those connected risks. Enter a public GitHub URL or an allowed local repository path and start a read-only scan. The app profiles the stack and files, then runs two diagnostic engines. TestPilot maps tests to source code and API routes, highlights missing or incomplete behavior checks, and can produce starter test skeletons. ConfigDoctor compares code, environment templates, containers, CI, runtime metadata, and documentation to find inconsistencies. The dashboard groups findings by severity and confidence, shows the supporting file evidence, and recommends a specific next step. It also provides a health score and downloadable Markdown and JSON reports. The distinction is evidence quality. Doctor Dev uses an actual coverage percentage only when it finds a supported coverage artifact. Otherwise it labels the result as evidence based or unavailable instead of inventing a percentage. Its included demo repository contains deliberate testing and configuration gaps for a repeatable walkthrough, while the same analysis pipeline also accepts other repositories. It is designed for maintainers, reviewers, and new contributors who need to decide what to fix before a release or handoff. The prototype currently has its deepest analysis for JavaScript and TypeScript, with an initial Python adapter. Generated tests are starting points for developer review.
27 Sep 2026
