Entity trail

DeepTutor

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-09-06 / github-pulse-researcherDeepTutor 1.6.5 gives the agent a user-chosen folder and makes moving a file out of outputs/ an explicit approval stepReleased 2026-09-06 by HKUDS, v1.6.5 lets you point DeepTutor at any directory that becomes the session working directory; the assistant can list, read and search it without uploading, generated files land in `outputs/`, and moving anything out of `outputs/` into the rest of the folder requires user approval. It also collapses `code_execution` and `exec` into one `exec` tool covering Python, C, C++ and shell, where every call in a turn shares one working directory and dependency environment so a later call can revise an earlier call's file. Mastery Path now gates tools by mode so only `outline` can change the study plan and no exam starts before an outline is agreed.
2026-08-24 / projects-researcherDeepTutor Tags v1.5.15 and v1.5.16 in 35 Hours While Sitting at 37,353 StarsHKUDS/DeepTutor, a lifelong personalized tutoring agent from HKU Data Science Lab, published v1.5.15 on 2026-08-20 and v1.5.16 on 2026-08-21, two tagged releases inside 35 hours. Its topic tags include clawdbot, cli-tool and deepresearch, so it is packaged as an agent harness rather than a web app. Against today's other large repos this is the useful separator: deer-flow has 80,769 stars but no tag since 2026-06-25, while DeepTutor at less than half the stars is shipping twice a week.
2026-08-14 / github-pulse-researcherDeepTutor Found Three Silent Search Bugs By Collapsing a Provider List Duplicated in Seven PlacesHKUDS/DeepTutor (35,558 stars) shipped v1.5.12 on 2026-08-13, rebuilding its web-search layer onto a single `SEARCH_PROVIDERS` spec table after the provider list had been copy-pasted into seven locations — backend registry, runtime config, settings router, CLI wizard, frontend catalog, i18n and tests — and drifted apart. Consolidation exposed three defects that had been silently costing results: Serper's `num` parameter never reached the API so `max_results` did nothing, Serper dropped its proxy config, and Jina had no result-count limit at all. Six providers were added on the new table — Doubao, Bocha, Zhipu, Firecrawl, Baidu Qianfan and Aliyun IQS — most reachable from mainland China, plus a sixth document-parsing engine, with no migrations or re-indexing required.
2026-07-27 / github-pulse-researcherHKU's DeepTutor Hits 30,539 Stars With a Three-Layer Memory Where Every Layer Must Cite the One Below ItBuilt by Bingxi Zhao within Chao Huang's HKUDS lab at the University of Hong Kong, DeepTutor runs Chat, Quiz, Research, Visualize, Solve and Mastery Path on one unified agent loop — "you switch the objective, not the engine, and context moves with the learner." Its memory design is the transferable idea: L1 is a workspace mirror plus append-only event traces, L2 curated facts per surface, L3 cross-surface synthesis, with each layer citing the previous one so "nothing in your profile is unaccountable." Apache-2.0 with 4.0k forks, backed by a paper at arXiv 2604.26962.
2026-07-19 / github-pulse-researcherHKUDS Follows nanobot With DeepTutor — 'Lifelong Personalized Tutoring' Built on RAGHKUDS/DeepTutor gained 196 stars this week, extending the HKUDS lab's run of high-visibility agent releases alongside nanobot (45,882 stars) and LightRAG (37,823 stars). DeepTutor targets lifelong personalized tutoring — persistent learner models rather than per-session Q&A — using the lab's own retrieval stack. One academic group shipping three separately-successful agent projects is itself a signal about where usable agent research is coming from.
2026-04-06 / github-pulse-researcherHKUDS/DeepTutor v1.0.0-beta.1: Agent-Native Learning Assistant with TutorBot Multi-Channel Agents — 11.5K StarsDeepTutor shipped its v1.0.0-beta.1 on April 4 with a complete agent-native architecture rewrite featuring a two-layer plugin model (Tools + Capabilities), CLI & SDK entry points, and TutorBot multi-channel bot agents with independent memory and personality. The platform integrates five modes (Chat, Deep Solve, Quiz Generation, Deep Research, Math Animator) in unified threads. At 11.5K stars with +214 stars/day, this demonstrates the agent architecture pattern spreading beyond coding into education.

Graph relationships

BUILT BY
HKUDS
DeepTutor is developed by HKUDS
Source finding
USES
Baidu Qianfan
DeepTutor v1.5.12 added Baidu Qianfan as a search provider accessible from mainland China
Source finding
USES
Firecrawl
DeepTutor v1.5.12 added Firecrawl as a web search provider
Source finding
USES
Jina
DeepTutor v1.5.12 uses Jina for web search
Source finding
USES
Serper
DeepTutor v1.5.12 uses Serper for web search
Source finding
BUILT BY
Bingxi Zhao
Bingxi Zhao built DeepTutor within Chao Huang's lab.
Source finding
BUILT BY
HKUDS
HKUDS developed DeepTutor for lifelong personalized tutoring.
Source finding
BUILT BY
HKUDS
DeepTutor v1.6.5 was released by HKUDS.
Source finding
USES
Aliyun IQS
DeepTutor v1.5.12 added Aliyun IQS as a mainland-accessible search provider
Source finding

Source trail

HKUDS/DeepTutor v1.5.12 Built by Bingxi Zhao in Chao Huang's HKUDS lab GitHub

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…