Foster Ledger Shift Notes
Working prototype · 3–5 minute recorded presentation · Source and tests
Problem and intended users
Foster homes and shelter volunteers need to carry small responsibilities between shifts. A board may show tasks without making the next owner or the unfinished handoff obvious. This is a design hypothesis awaiting coordinator interviews, not a claim based on completed user research.
Solution
Shift Notes combines my existing animal-care task board with a new plain-text handoff. Volunteers can add tasks, claim them, mark them complete and reopen them. The new report orders tasks by due time, identifies past-due unfinished work and produces a readable snapshot for someone who does not use the app. It includes unfinished work by default, with controls to include completed tasks or omit volunteer owner fields.
The preview and downloaded text include a timestamp, explicit UTC dates and scope. Omitting owner fields does not anonymize text elsewhere; users must review descriptions before sharing. The note is a snapshot and does not update when the board changes.
Implementation and plausibility
The prototype uses browser JavaScript, HTML and CSS. Local storage preserves the board on that device. The task model validates dates, state transitions, required owners and imported records. The new report module creates plain text from the validated board and normalizes embedded line breaks to keep the report structure readable. There is no server synchronization or authentication, so a shared-device workflow and manual file handoff are the current practical limits.
Prior projects, sources and AI
Practice to Create explicitly permits previous projects with disclosure. This extends my Foster Ledger AnimalHack entry, itself adapted from my SlothDO source. The original board and history are preserved. New work for this event includes the readable report, scope and owner-field controls, preview, report tests, feedback plan and narrated presentation.
OpenAI Codex assisted with implementation, testing and documentation. Google Cloud Chirp 3 HD generated stock-voice narration over an actual screen recording of the prototype. No voice cloning was used. The live interface and all code shown in the presentation are publicly linked above; the narration script is in .autohack/NARRATION.txt.
Validation and feedback request
Seven model/report tests pass, including ownership transitions, malformed imports, completed-task filtering, owner-field omission and multiline text. Browser checks verify the report preview, actual plain-text download and mobile layout.
The proposed evaluation asks whether a new volunteer can identify the next unclaimed task, whether the note has enough context, and whether a shared-device workflow fits the organization. A consented usability session with fictional animals and coordinator interviews would test that hypothesis. No pilot, adoption, improved care or commercial outcome is claimed.
Example animals are fictional. Self-entered aliases are not authenticated identities, and a local activity list is not proof of completed care. The app provides no veterinary advice, medication instructions or emergency monitoring. Amrit Lahari, BITS Pilani, is the sole entrant.
Built With
- css
- google-cloud-text-to-speech
- html
- javascript
- node.js
- openai-codex
- playwright

Log in or sign up for Devpost to join the conversation.