Choose theme
AI Visuals
DeepTutor
DeepTutor is a learning workspace from HKUDS for asking questions about study materials and practising what you have learned. It connects tutoring, knowledge bases and learner memory through a web app and CLI.
Mastery Path can save a graded question to the Question Bank with your answer, the reference answer and an explanation, so a missed question becomes material you can revisit. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
Study materials and practice in one workspace
Bring learning materials into a knowledge base, ask questions and generate practice. The technical report describes tutoring grounded in retrieved sources alongside question generation calibrated by difficulty.
Why it stands out
Learner memory you can examine
Memory separates live workspace entities, summaries for individual learning surfaces and a synthesis across surfaces. The workbench lets you inspect and edit what personalization retains rather than treating every remembered fact as fixed.
Availability
A local app with configured model providers
The main Python package includes the web app and CLI. You still configure a model endpoint and credentials; hosting the app locally does not by itself make inference local or offline.
Why it matters
What makes it useful
A diagram in your textbook needs more than text retrieval. DeepTutor’s LlamaIndex pipeline can retain extracted figures, captions and source locations. A vision-capable answer model can receive retrieved pixels; a text-only model receives captions and context. This requires an image-emitting parser, and older material must be reindexed. Separately configured image and video models can become chat generation tools with inline previews and downloads; Visualize also provides charts and diagrams. Those creation tools need their own enabled configuration.
What to know
Where it fits
Mastery Path organizes a topic into modules and knowledge points, with assessed gates governing progression. A learner can also explicitly mark a point mastered; the workspace distinguishes that declaration from assessed evidence. For a scripted research-to-quiz workflow, reuse the CLI session ID to retain the conversation and repeat any needed --kb, --tool and --language selections: omitted flags replace those preferences with empty selections or the default language.
Notable points
What stands out
Separate workspaces keep learning materials, conversations, progress and caches apart, but Memory, settings and credentials remain shared at the account level. Switching workspace is therefore not a way to start with a separate learner memory.
Before using
What to review
The packaged web app needs Python 3.11–3.14 and Node.js 20 or later. The CLI-only distribution has a separate source-install path.
Choose the model and any embedding or search providers with the materials you intend to submit in mind. A local interface does not establish where those configured providers process requests.
Check explanations against the original material before using them for assessed work. Source grounding and stored reference answers do not establish educational accuracy.
If a tutoring session will not start, compare deeptutor doctor with deeptutor doctor --online. The first checks local runtime readiness; the second also sends a small request to the configured provider. This separates local setup from provider-response problems; neither report tests lesson accuracy.
Reader fit
Who may find it relevant
Learners who want to return to missed questions and continue studying their own materials.
Developers who need to chain tutoring capabilities through structured CLI output.
People seeking a ready-to-use hosted tutor still need to arrange the app and model-provider setup described here.
Editorial note
Why LifeHubber lists it
DeepTutor is useful to examine when study needs to continue along a named route over several sessions. Mastery Path keeps the topic’s conversations with its map and current waypoint, while completed points can return for spaced review when due. This gives a learner somewhere to resume beyond a collection of saved questions or personalization notes. The progression record helps organize practice; it still needs the learner’s judgment and checks against the original material.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
More in AI Agents
Keep browsing this category
Explore more AI agent projects.
Paperclip
paperclipai/paperclip
A self-hosted server and dashboard for coordinating agent teams through companies, goals, roles, issues, heartbeats, budgets, approvals, and persistent activity records.
Magpie
yetone/magpie
A local model gateway and configuration app for supported coding agents, with provider choices, routing groups, saved profiles, and usage views. It changes agent settings and can discover existing subscription sign-ins. Requests go to the chosen provider, and released builds send a daily user-count event unless disabled.
Vane
ItzCrazyKns/Vane
A self-hostable AI answering engine for configurable search-style workflows, with local and cloud model providers, SearxNG-backed web search, cited sources, file uploads, and Docker setup paths.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
