Skip to content
BranchesTags

Repository files navigation

Velo - your Canvas co-pilot

Flutter Dart Material 3 Canvas LMS API Groq Deployed on Vercel Made with AI License: MIT

Velo is a distraction-free, local-first client for the Canvas LMS that helps university students see, plan and submit their whole workload in one place.
Landing page velo-landing-page-delta.vercel.app
Live app velo-canvas-copilot.vercel.app
Demo video Watch on Google Drive
Course Applications Development and Emerging Technologies (6ADET), Holy Angel University
Author Sean Rhani J. Dela Cruz

Contents

Try it

Velo is published on the web, so there is nothing to install.

  1. Open the live app.
  2. In Canvas, go to Account > Settings > New Access Token and generate a token.
  3. Paste the token on Velo's sign-in screen.

The live app is connected to Holy Angel University's Canvas, so it needs a HAU Canvas account. On a phone, use your browser's Add to Home Screen to install it like an app.

Screenshots

Log in To do Courses
Log in To do Courses
ADET course Announcements Modules
ADET course Announcements Modules
Grades Assignments Planner
Grades Assignments Planner
AI Assistant Inbox Side drawer
AI Assistant Inbox Side drawer
Account & Settings
Account and Settings

What it does

Feature What it does
Sign in Signs in with your own Canvas access token. The token is checked against Canvas and kept only on your device.
To do One feed of every unsubmitted task across all your courses, sorted by deadline, with a summary of what is overdue, due today and due this week. Each card carries its course image.
Courses Every active course with its Canvas image or colour and its count of active tasks. Each course opens to its announcements, modules, grades and assignments.
Announcements Course announcements with their original formatting, clickable links and downloadable attachments.
Modules The course's modules and items, with pages and assignments readable in the app and previous/next navigation.
Grades Your current course grade and the score for each assignment, with late, missing and excused flags.
Assignments and submissions Assignment instructions from Canvas, and real submissions as a file upload, a text entry or a website URL. The comment thread with your instructor is shown and you can post to it.
Planner Breaks assignments into daily milestones. Auto-Plan asks the AI to spread upcoming work across the days before each deadline, and every step stays editable.
AI Assistant A chat that answers from your real Canvas data: tasks, grades, announcements, messages and assignment details. Replies link straight to the task or thread in the app.
Inbox Canvas conversations and course announcements together. Read, reply, compose, archive and delete, with an unread badge in the navigation.
Offline mode Everything you have opened is cached on the device. Screens load from the cache first and refresh in the background, and offline mode shows saved data with the last sync time.
Account & Settings Your Canvas profile and bio, light and dark themes, deadline reminders, and a shortcut to Campus++.

Built with

Layer Tools
Framework Flutter (Dart), Material 3
State provider
Storage shared_preferences (local-first cache and offline fallback)
Network http
Backend Canvas LMS REST API
AI Groq chat completions API, with function calling
Routing go_router
UI flutter_widget_from_html, flutter_markdown_plus, fl_chart
Utilities file_picker, url_launcher, intl, timezone, flutter_local_notifications, flutter_dotenv
Hosting Vercel

Privacy and secrets

  • Personal data: Velo is local-first. Your Canvas token, profile and cached course data are stored on your device with shared_preferences, and Velo has no database of its own. Signing out clears them.
  • What leaves the device: requests go to Canvas, through the hosting proxy on the web build. When you use the AI Assistant or Auto-Plan, the Canvas data needed to answer (such as task names, deadlines, grades or messages) is sent to Groq.
  • Secrets: API keys are read from a .env file that is not in this repository.
  • Sample data: the screenshots in this repository are checked for personal information before they are added.

Project documentation

Document What is in it
Proposal the problem, the users, the scope
Mockup and wireframes what it looks like, and the screen flow
Design system colors, type, spacing, components
Weekly reports what happened each week
Demo video the recording and what it shows
Security and privacy the checklist, filled in
Presentation the week 3 presentation video
AI usage how AI was used, where it was wrong, and who wrote what
Documentation guide What is in it
Week 1 the project documentation guide for week 1
Week 2 the project documentation guide for week 2

Status and what is next

What works:

  • Sign-in with a Canvas access token, and sign-out that clears saved data.
  • The To do feed, Courses, Announcements, Modules, Grades and Assignments, all on live Canvas data.
  • Real assignment submissions (file, text entry, URL) and submission comments.
  • The Planner with AI Auto-Plan, and the AI Assistant with function calling.
  • The Inbox: read, reply, compose, archive and delete.
  • Local caching, offline mode, pull to refresh, and light and dark themes.
  • The web deployment, installable as a PWA.

Known limits:

  • Sign-in uses a manually generated access token, not Canvas OAuth.
  • Media-recording submissions open in Canvas; they are not sent from the app.
  • A submission carries one file.
  • Deadline reminders work on mobile builds, not on the web build.

What is next:

  • Canvas OAuth sign-in in place of manual tokens.
  • Media-recording and multi-file submissions.
  • Notifications on the web build.

Credits

  • Packages: see pubspec.yaml
  • Icons: Flutter's Material Icons
  • Backend: the official Canvas LMS REST API
  • AI model hosting: Groq

AI use

This project was built with AI assistance. Assistant used and how much: TODO. The full record, with a commit link for each use, the cases where the AI was wrong, and the parts written by hand, is in AI-USAGE.md.

Licence

MIT, see LICENSE.

About

Velo is a distraction-free, local-first mobile client for the Canvas LMS designed to help university students manage heavy project workloads. It consolidates active tasks across all subjects into a single, chronologically sorted mobile feed and features a conversational AI Assistant for querying syllabus details and deadlines in natural language.

velo-canvas-copilot.vercel.app

Topics

aiai-agentai-assistantsapiapi-integrationcanvas-lmsdartflutterminimalistmobile-developmentrest-api

Resources

Readme
MIT license
Activity

Stars

0 stars

Watchers

0 watching

Forks

0 forks
Report repository

Contributors

Languages

Generated from HAU-6ADET/final-project-template