HKUDS

Does Graph RAG Actually Help? I Measured It on Postgres and TypeScript Graph RAG is usually described with one claim: vector search misses answers that sit behind a chain of facts, and a knowledge graph fixes that. That is plausible. It is also expensive to adopt, because every document goes through an LLM before it is searchable. So I built a small version on Postgres, ran it against a hybrid search baseline on 100 multi-hop questions, and counted what it cost. The short result: graph retrieval helped, mostly when combined with hybrid search, and the one-time indexing bill was about twelve times the token cost of the whole hybrid evaluation. The details matter more than the headline, so this article walks through the setup, the numbers with their uncertainty, the questions where the graph made things worse, and what the experiment cannot tell you. All code and data are in aiengineering-book/graph-rag-experiment. You can skip the expensive build step and reproduce the results with a 0.8 MB file. --- What Was Tested Four systems answer the same questions with the same answer prompt, so the only difference between them is which passages the model sees: | System | What it retrieves | |---|---| | Closed-book | Nothing. The model answers from what it already knows (a control). | | Hybrid | Vector search and Postgres full-text search (ts_rank), merged by reciprocal rank fusion. Top 5 passages. No reranker. A BM25 version is tested later. | | Graph | Passages found by walking a knowledge graph from the entities in the question. Top 5 passages. | | Graph + hybrid | The two ranked lists merged by reciprocal rank fusion. Top 5 passages. | The questions come from MuSiQue (CC BY 4.0), a benchmark of questions that need two to four reasoning steps across different passages. For example: > Where does Ulysses Kay's birth city hold NASCAR races? Answering it takes two passages: one that says Ulysses Kay was born in Tucson, and one that says what races Tucson hosts. The question names only the first. I sampled 100 questions from the dev set with a fixed seed. Each question ships with 20 passages (19 for one of them), a few relevant and the rest distractors. I pooled all of them into one corpus of 1,847 passages, so every system searches the same collection, as it would in a real deployment. The stack: - Postgres 16 with pgvector. Vectors, full-text search and the graph all live here. There is no graph database. - nomic-embed-text, running locally through Ollama, for all embeddings (768 dimensions). - DeepSeek (deepseek-flash, thinking disabled) for graph extraction, for finding entities in a question, and for answering. Any OpenAI-compatible chat API would work for the same roles. Scoring is exact match and token F1 against the gold answer and its aliases, plus the share of each question's supporting passages that a system actually retrieved. Exact match is strict ("The Mother (Tracy McConnell)" does not match "Tracy McConnell"), so F1 is reported too. --- How the Graph Is Built and Searched Indexing adds one step to what a hybrid index does. An LLM reads each passage and returns the entities in it and the relations between them as JSON: { "entities": ["Ulysses Kay", "Tucson, Arizona"], "relations": [ { "source": "Ulysses Kay", "relation": "born in", "target": "Tucson, Arizona" } ] } The results go into three tables next to the passages. Entities are merged by normalized name, and each relation remembers which passage it came from: CREATE TABLE entities (id serial PRIMARY KEY, name text, key text UNIQUE, embedding vector(768)); CREATE TABLE mentions (entity_id int REFERENCES entities, chunk_id int REFERENCES chunks, PRIMARY KEY (entity_id, chunk_id)); CREATE TABLE relations (id serial PRIMARY KEY, src int REFERENCES entities, dst int REFERENCES entities, label text, chunk_id int REFERENCES chunks); Searching the graph (what GraphRAG calls local search) has four steps: 1. Ask the LLM for the named entities in the question. 2. Embed each name and find the nearest entity in the entities table. These are the seeds. 3. Walk the relations table outward from the seeds for two hops. 4. Score each passage by the entities it mentions, weighting each by 1 / (1 + hops from a seed), and keep the top 5. For the Ulysses Kay question, the seed is "Ulysses Kay". One hop reaches "Tucson, Arizona", so the passage about Tucson is retrieved even though the question never mentions it. This is a deliberately small graph. It has no community detection and no community summaries, which Microsoft's GraphRAG uses for questions about a whole corpus. That matters for what the results can and cannot say, and it is covered at the end. --- Results | System | Exact match (95% interval) | F1 | Supporting passages retrieved | Exact match vs hybrid, same questions | |---|---|---|---|---| | Closed-book | 10% | 0.19 | n/a | n/a | | Hybrid | 28% [19, 37] | 0.42 | 51% | baseline | | Graph | 35% [26, 45] | 0.48 | 61% | +7 points [-3, 17]; right where hybrid was wrong 18 times, the reverse 11 | | Graph + hybrid | 42% [33, 52] | 0.53 | 68% | +14 points [5, 24]; 20 times against 6 | The intervals are 95% bootstrap intervals over the 100 questions. The last column compares systems on the same questions, which is more informative than comparing two percentages. What the table supports: - Retrieval helps at all. Without it, the model answers 10% of these questions. Hybrid search triples that. - Graph + hybrid beat hybrid. The interval for the difference, [5, 24] points, excludes zero. - Graph alone did not clearly beat hybrid. Its interval, [-3, 17], includes zero. Seven points is the best estimate, but 100 questions cannot separate it from noise. - Hybrid misses half the evidence. It retrieved 51% of the supporting passages. That gap is the room a graph has to work with. --- Where the Graph Helped The 18 questions the graph answered and hybrid did not share a shape: the answer sits behind an entity the question does not name. For the Ulysses Kay question, hybrid search matched the words "NASCAR races" and returned passages about Richmond, Virginia, a NASCAR TV show and two NASCAR drivers, then answered "Richmond International Raceway". The graph started from "Ulysses Kay", followed the birthplace relation to Tucson, and answered "Tucson Raceway Park". The same pattern appeared for "Who is the child of the actor who is a cast member of Love in Pawn?", for "In which county was James Finch born?", and for "Who was the spouse of the leading lady in Gone With the Wind?". In each case the second passage is reachable by following a relation from a named entity, and not by similarity to the question's words. --- Where It Hurt The graph answered 11 questions worse than hybrid (two of them are the same question, which appears twice in the sample). Two causes showed up. The question has no entity to start from. "What was the percentage that the Dow Jones fell when the US had the highest unemployment rate?" contains no usable named entity, so the graph search had no seeds, returned unrelated passages, and the model answered "I don't know". Hybrid found it from the words alone. The seed is a hub. For "In what part of Florida is Tom Denney's birthplace located?", the question names "Florida" as well as "Tom Denney". "Florida" has 41 neighbors in the graph, and walking out from it pulled in passages about other Florida cities, which crowded out the correct one. The second cause suggested a fix: do not walk outward through entities with many neighbors. The median entity in this graph has one neighbor and the 99th percentile has 13, so I set the cutoff at 20, which catches 43 hub entities including "Florida". That fixed the Tom Denney question, and it raised graph + hybrid from 42% to 48%. There is a catch, and it affects how much to trust that number. I chose the cutoff after reading the failures in this same sample of 100 questions. Even with the paired comparison, a gain from a setting tuned on the test set is optimistic, and its interval ([0, 12] points over graph + hybrid without the cap) barely clears zero. The Tom Denney question also appears twice in the sample under different ids, so that fix is counted twice. Treat the hub cap as a hypothesis to confirm on fresh questions, not as a result. --- What Else Changed the Outcome Each of these reuses the same graph, so there is no rebuild: | Variation | Exact match | Supporting passages retrieved | |---|---|---| | Graph + hybrid, 2 hops (default) | 42% | 68% | | Graph + hybrid, 1 hop | 36% | 64% | | Hybrid, 8 passages instead of 5 | 33% | 60% | | Graph + hybrid, 8 passages | 48% | 76% | | Hybrid, BM25 for the keyword half | 27% | 59% | | Graph + hybrid, BM25 for the keyword half | 38% | 68% | - The second hop matters. With one hop, graph + hybrid falls to 36%, and its advantage over hybrid no longer clears zero. Most of the benefit comes from reaching passages two steps away. - A stronger keyword search did not rescue the baseline. The keyword half of the hybrid search above is Postgres's built-in full-text ranking, which is not BM25. Swapping in BM25 (same parameters as the one in the book: k1 = 1.5, b = 0.75) raised the share of supporting passages hybrid retrieved from 51% to 59%, but exact match stayed at 27%. Graph + hybrid with BM25 scored 38%, which is 11 points above hybrid with BM25 [3, 20], winning 16 questions to 5. That is a little smaller than the 14 points against the original baseline, and the interval still excludes zero. My guess for why the better retrieval did not help: an answer needs every hop's passage, and BM25 mostly found more of the first-hop ones. - The gain is not just more context. Giving hybrid search 8 passages lifts it from 28% to 33%. Graph + hybrid with 8 passages reaches 48%, ahead of hybrid with the same 8 passages by 15 points [7, 23], on 17 questions against 2. That extra context costs about 50% more prompt tokens (121,000 against 80,000 per 100 questions). --- What It Costs | Item | Cost | |---|---| | Building the graph, once | 1,827 LLM calls, about 964,000 tokens (450,000 prompt, 514,000 completion), about 14 minutes at 4 concurrent requests | | Evaluating hybrid on 100 questions | about 80,000 prompt tokens | | Evaluating graph + hybrid on 100 questions | about 82,000 prompt tokens, plus one extra LLM call per question to find the question's entities | | Updating one edited passage, graph | about 255 prompt and 306 completion tokens, about 2 seconds | | Updating one edited passage, hybrid index | about 70 to 90 ms, no LLM call | The build is about twelve times the token cost of the entire hybrid evaluation, and nearly all of the experiment's spend. At query time the graph adds one short LLM call and about the same prompt size as hybrid, so serving cost is modest. The cost lives in indexing. Two costs are not visible in those numbers: - Extraction fails sometimes. 57 of 1,847 passages (3%) returned JSON that did not parse, so they have no graph entries at all. - Entity merging is crude. Entities merge by normalized name, so "Future Ted" and "Future Ted Mosby" ended up as separate entities. A split like that can break a path. Better entity resolution would cost more. The update figure applies to this small graph, where re-extracting one passage is enough. A system that also keeps community summaries has to regenerate those after changes, which this experiment did not measure. --- What This Does Not Show - One run on 100 questions, with one model. The intervals are wide. A different sample, or a different answering and extraction model, could move the numbers by several points. Stronger models might narrow the gap by making hybrid search better at using the passages it finds, or widen it by extracting a better graph. I did not test either. - Only multi-hop questions. The benchmark has no single-fact questions, so this says nothing about whether the graph helps or hurts them. It also has no whole-corpus questions of the "which risks recur across all these contracts" kind. - No community summaries. Those are what GraphRAG's global search relies on. This experiment tests the local half of graph retrieval only. Whether a graph helps on whole-corpus questions needs a different system and a different corpus. - No reranker. I tested BM25 against the built-in full-text ranking, but not a reranking step (a cross-encoder that rescores the top candidates). It could lift the baseline further. It only reorders passages that were already retrieved, so I would expect it to help less with passages found through a bridge entity, but I did not test that. - No comparison with multi-step retrieval. An agent that rewrites its query after each retrieval round can also follow a chain of facts, with no extra index. I did not compare against it. - Possible training-data overlap. The model answers 10% of the questions with no retrieval at all, which is modest but not zero. The control run is there so you can see it. --- When It Is Worth Trying The experiment points to a narrow case. A graph is worth evaluating when: - a large share of your questions need facts from several documents joined through an entity the question does not name; - your documents describe entities and relations that stay fairly stable, because every change means re-extraction; - you can afford an indexing bill that scales with corpus size and can measure the gain against your current hybrid search first. If most of your questions are single-fact lookups, hybrid search with reranking is the cheaper path, and this experiment gives you no reason to change it. And before building a graph, measure how many of your questions are multi-hop at all. If that share is small, the cost is unlikely to pay back. --- Reproduce It Clone the repo and start Postgres with pgvector (docker compose up -d). Install Ollama and run ollama pull nomic-embed-text. Then: npm install cp .env.example .env # add an OpenAI-compatible API key for answering npm run import-graph # loads the finished graph, about 2 minutes, no LLM calls npm run eval -- --system hybrid npm run eval -- --system graph+hybrid npm run eval -- --system hybrid --keyword bm25 # the BM25 variants npm run report -- results/hybrid-100.json results/graph-hybrid-100.json import-graph copies the graph exactly but recomputes the embeddings locally. Embeddings come out slightly different on each import, so on the first 30 questions a few near-tied searches returned a different top 5, and exact match moved by two questions. Expect your numbers within a few points of the tables above, not identical to them. To try a different answering model, the repo also ships the exact prompts each system sent (prompts/*.jsonl, retrieval already done). npm run run-prompts sends them to any OpenAI-compatible API, and npm run score-answers scores the result, with no database or Ollama needed. I only tested DeepSeek, so another provider may need small client changes. To rebuild the graph yourself, download MuSiQue and draw the sample with scripts/sample-musique.py (the steps are in data/README.md), then run npm run ingest and npm run build-graph in place of import-graph. The build is the 964,000-token step. --- Further Reading Hybrid search with BM25 and reranking, the evaluation method, and a section on deciding when graph-based retrieval is worth adopting are covered in Chapters 11 and 12 of 《AI Engineering with TypeScript — A Comprehensive Guide to Building AI Agents》 at Leanpub. The book's companion code is at github.com/aiengineering-book/code. The experiment code, the finished graph and the per-question results are in aiengineering-book/graph-rag-experiment. The questions come from MuSiQue: Trivedi et al., "MuSiQue: Multihop Questions via Single-hop Question Composition", 2022 (repository, CC BY 4.0). Microsoft's GraphRAG and HKUDS's LightRAG are the two open-source implementations that include community summaries and global search.

DeepTutorをコードから読む：マルチエンジンRAGの構成と静的セキュリティレビュー TL;DR HKUDS/DeepTutor は香港大学 Data Intelligence Lab（HKUDS）による学習特化のエージェントアプリ。2026-10-30 時点で ★50,686（当日 +1,930）、Apache-2.0 知識ベースは LlamaIndex / PageIndex / GraphRAG / LightRAG など複数の RAG パイプラインを差し替え可能な構成 コミット 1ea7df1554 を実行せずに静的レビューした。自動スコアは 23/100（RED）だったが、high 検出は精査するとほぼ誤検知。ただし v1.4.x に CVSS 最大 9.8 の CVE 群があり、現行 v1.7.0 を使う前提で「注意つき」と判断した ! 本記事はリポジトリの静的な読解と公開情報に基づきます。コードの実行・動的解析はしていません。 概要 DeepTutor は「AI チューター」を自分の環境で動かすための OSS です。バックエンドは Python（3.11–3.14）、フロントは Next.js で、deeptutor start が同梱の Next.js standalone サーバを起動する構成です。README によると、会話・出題・小テスト・調査・可視化・解答・マスタリーパス・精読・音声まとめといったモードが同じ会話コンテキストを共有します。 リポジトリの指標（スキャン時点）: 項目 値 初回コミット / 最終コミット 2025-12-28 / 2026-10-29 コミット数 3,155（直近30日 738） 作者数 278（直近90日アクティブ 164） トップ作者比率 0.39 CI ワークフロー 5 テスト あり 1年弱でこの規模という、かなり速いペースで開発されているプロジェクトです。 アーキテクチャ：retrieval まわり deeptutor/ 配下は agents, capabilities, knowledge, services, tools, runtime などに分かれています。RAG の中心は deeptutor/services/rag/ です。 deeptutor/services/rag/ ├── factory.py / service.py # RAGService とパイプライン選択 ├── query_rewrite.py # 検索前のクエリ書き換え ├── smart_retriever.py # マルチクエリ検索と集約 ├── index_versioning.py # KB インデックスのバージョン管理 ├── embedding_binding.py ほか # 埋め込みプロバイダの紐付け └── pipelines/ ├── llamaindex/ pageindex/ graphrag/ ├── lightrag/ lightrag_server/ └── weknora/ ima/ kiwix/ 1. クエリ書き換え（fail-open） query_rewrite.py は、会話中の指示語を含む質問（「それってどういう意味？」など）を、LLM を 1 回呼んで自己完結したキーワード寄りの検索クエリに書き換えます。docstring に fail-open と明記されていて、LLM のエラー・タイムアウト・空出力のときは元のクエリにフォールバックします。書き換えに失敗しても検索自体は止まりません。現時点で実装されている戦略は standalone だけで、multi_query は今後の PR 用に予約されています。 2. マルチクエリ検索 SmartRetriever は、コンテキストからクエリのバリエーションを生成し（既定 max_queries=3、ヒントが渡されればそれを使う）、asyncio.gather(..., return_exceptions=True) で並列に検索して結果を集約します。1 本のクエリが例外を出しても、残りの結果は生かされる作りです。 3. パイプラインとインデックスのバージョン管理 KB はバージョン付きの version-N ディレクトリとして管理されます。たとえば PageIndex は埋め込みを使わないので、ローカルに置くのは「取り込んだファイル → ホスト側の doc_id」の対応表（pageindex_docs.json）だけです。それでも同じ index_versioning を使うことで、UI 上はほかの KB と同じように扱えるようになっています。プロバイダ定数には pageindex（クラウド）と pageindex-oss があります。 ! PageIndex（クラウド版）のように、ドキュメント自体が外部サービスに置かれるパイプラインもあります。教材に機密が含まれる場合は、どのパイプラインを選ぶかがデータの出口の選択になります。 セットアップ上の注意（API キー・コスト） お試しなら Docker イメージを使えます（README 記載）。 docker run --rm --name deeptutor \ -p 127.0.0.1:3782:3782 \ -v deeptutor-data:/app/data \ ghcr.io/hkuds/deeptutor:latest 127.0.0.1 にバインドしている点に注目してください。.env.example のコメントでも、ホスト側ポート（8001 / 3782 / 8090）はすべて loopback 前提とされています。起動したら Settings → Providers でモデルプロバイダを設定します。 pip でインストールする場合、1.7.0 からモジュール構成が変わっています。 pip install -U "deeptutor[full]" # Web アプリ + CLI 素の deeptutor は core のみ README には、旧 deeptutor-cli パッケージが入っている場合は先にそれをアンインストールするよう書かれています。同じトランザクションでアップグレードすると、新しい core が消えることがあるためです。 ! LLM・埋め込み・検索は設定したプロバイダへ送信されます。各社の API キーが必要で、従量課金になる可能性があります。特にクエリ書き換えとマルチクエリ検索は、1 回の質問で複数回 LLM / 検索を呼ぶ設計です 個人情報や機密資料は入れないでください deeptutor/services/llm/telemetry.py はローカルにログを出すだけで、外部には送信していませんでした（ただし全通信先を網羅的に確認したわけではありません） 静的セキュリティレビュー 対象はコミット 1ea7df155447f4caf0e3abc86b3ff5395479cd4a（2026-10-29 01:36 JST）、スキャン日時は 2026-10-30 15:17 JST です。 自動スコア：23 / 100（RED） 区分 件数 減点 high 15 -60 medium 1 -8 low 10 -6 SECURITY.md なし — -3 数字だけ見ると「紹介しない」側ですが、検出内容を 1 件ずつ見ていきます。 検出の精査（ほぼ誤検知） CREDENTIAL_PATHS（13件）：chat_hints.py, security_guard.py, routers/book.py, フロントの ChatStateAdapter.tsx など。検出された文字列はすべて local state で、UI 状態やローカルストレージの意味で使われていました。たとえば pageindex/storage.py:4 は、前述の「The only local state is a lightweight manifest…」という docstring です。SSH 鍵やブラウザの認証情報に触れるコードではありません → 誤検知。 HIDDEN_UNICODE（2件）：tests/runtime/providers/test_text.py と tests/services/cli_apps/test_provider.py。ゼロ幅スペースや RLO を除去するサニタイザのテスト入力です → 誤検知。 VERY_LONG_LINE（1件）：web/lib/brand-icons.generated.ts（11,747 文字）。生成されたアイコンデータです → 誤検知。 PIPE_TO_SHELL：deeptutor/services/cli_apps/models.py:10 は、レジストリにある curl … | bash を実行しないことを説明する docstring でした。残りは tests とリリースノート（ver1-5-7.md）だけです → 誤検知。 RAW_IP_URL：tests 内のみです。169.254.169.254（クラウドのメタデータエンドポイント）は SSRF 防止のテストケースでした。 Bandit B202（deeptutor/tools/tex_downloader.py:180）： if not is_within_directory(extract_dir, member_path): print(f"Suspicious file path in tar: {member.name}. Skipping.") continue yield member tar.extractall(extract_dir, members=safe_members(tar)) def _extract_zip(self, zip_path: Path, extract_dir: Path): with zipfile.ZipFile(zip_path, "r") as zip_file: zip_file.extractall(extract_dir) tar 側は safe_members でディレクトリ外を指すメンバーを除外しているので、実害は小さいと判断しました。一方 zip 側は同じ検証をしていません。軽微な注意点です。 エージェント向け指示書 フック（AGENT_HOOKS）や MCP 設定ファイルは見つかりませんでした。AGENTS.md / SKILL.md のほか、capabilities//prompts/ や skills/builtin//SKILL.md など多数のプロンプト用 Markdown があります。冒頭を中心に確認した範囲では、隠れた指示や外部送信命令はありませんでした。grep で探した範囲では不可視文字もありません。ただし全文を精読したわけではありません。 既知の CVE と修正バージョン ここが一番重要なポイントです。 CVE 影響 修正 CVE-2026-51876 / 51878 / 51879 / 51880 / 51881 v1.4.0。RCE・パストラバーサル・認可回避（CVSS 最大 9.8） v1.4.1（シェル実行を既定で無効化 など） CVE-2026-58168 1.4.10 未満。MCP ツールの認可回避 1.4.10 現行は v1.7.0 です。マルウェア混入やアカウント乗っ取りの報告は見つかりませんでした。 ! 1.4.10 未満は使わず、必ず最新版にアップデートしてください インターネットに公開しないでください（README の Docker 例も loopback バインドです） シェル実行や MCP ツールなど権限の強い機能は、必要なとき以外は有効にしないでください 衛生面 SECURITY.md・CODEOWNERS・dependabot がない：脆弱性の報告窓口が明文化されていません。過去に重大な CVE が出ている規模のプロジェクトとしては弱いところです CodeQL などのコードスキャンもありません。pull_request_target は使われていません Python 依存はピン留めされていないため、PyPI advisory との自動照合はできませんでした このレビューの限界 grep・Bandit・一部の目視による静的チェックと公開情報だけで判断しています。実行・動的解析はしていません 巨大なコードベースのすべてを読んだわけではなく、通信先も網羅的には確認していません 依存関係の脆弱性は未検証です 「安全である」という保証ではなく、「紹介してよいか」を判断する材料です まとめ DeepTutor は、RAG パイプラインを差し替えられる設計と、fail-open なクエリ書き換え・マルチクエリ検索を持つ、読みごたえのあるコードベースです。自動スキャンの RED 判定は大半が誤検知でした。一方で、過去の CVE 群と SECURITY.md がない点は実際のリスクです。最新版を使う・ローカル限定で動かす・強い権限の機能は無効にしておく、この 3 点を守れば試す価値は十分にあります。 参考： HKUDS/DeepTutor github.com/HKUDS/DeepTutor/releases/tag/v1.4.1 app.opencve.io

This repo gives AI agents a command line for Blender, FreeCAD, draw.io and even video games. 🔗 HKUDS/CLI-Anything Credits: HKUDS/CLI-Anything #clianything #python #aiagents

Hardening DeepTutor's Lifelong Knowledge Tracing in Production Tutoring Pipelines Integrating DeepTutor: What the Docs Don't Tell You When we evaluated HKUDS/DeepTutor (HKUDS/DeepTutor) for a production adaptive learning stack, the pitch was straightforward: lifelong personalized tutoring with persistent knowledge tracing. The reality? The framework's stateful architecture introduces failure boundaries you won't hit until multi-tenant load testing. DeepTutor aims to model long-term student progress by maintaining evolving knowledge states across sessions. Unlike ephemeral chatbots that treat every interaction as isolated, this system must reconcile historical performance data with real-time tutoring interactions. That persistence layer becomes your bottleneck. Architecture Breakdown: Where State Lives The core challenge in DeepTutor isn't the tutoring logic—it's the knowledge tracing pipeline. The system maintains: 1. Student Knowledge State: Vector embeddings representing concept mastery over time 2. Interaction History: Timestamped question-answer pairs with correctness labels 3. Adaptive Curriculum: Dynamic question selection based on current state estimates In our test topology, we ran DeepTutor behind a FastAPI gateway routing to multiple LLM backends through B-Lost's relay gateway (b-lost.com) for provider fallback. Configuration: import openai client = openai.OpenAI( base_url="https://b-lost.com/v1", api_key=os.getenv("BLOST_API_KEY") ) def generate_tutoring_response(student_state, question): prompt = build_adaptive_prompt(student_state, question) response = client.chat.completions.create( model="gpt-4o", messages=[{"role": "user", "content": prompt}], temperature=0.7 ) return response.choices[0].message.content The issue: DeepTutor's knowledge tracing runs inference on every interaction to update student embeddings. With 200ms average LLM latency plus 50-80ms for state vector updates, single-student sessions clocked 250-280ms total response time. Acceptable for demos, but not for classrooms with 30 concurrent learners. Session Isolation and State Collision The repository doesn't document multi-tenant isolation strategies. We discovered this the hard way when parallel sessions corrupted student state vectors. The culprit: # Naive global state management (anti-pattern) student_states = {} def update_knowledge_state(student_id, interaction): if student_id not in student_states: student_states[student_id] = initialize_state() # Race condition: concurrent updates to same student_id student_states[student_id] = recalculate_state( student_states[student_id], interaction ) Production fix required session-scoped state management with Redis-backed locking: # Terminal session diagnostic redis-cli KEYS "deeptutor:student:*" redis-cli GET "deeptutor:student:12345:state" redis-cli EXPIRE "deeptutor:student:12345:lock" 5 Migrating to Redis cut state collision errors from 12% to under 0.3% in concurrent load tests (simulated 50 students, 5 questions/min each). RAG Integration: When Retrieved Context Conflicts with Knowledge State DeepTutor's adaptive curriculum pulls from a static knowledge graph. We augmented it with RAG for dynamic content (current events, updated curriculum standards). Configuration: from langchain.vectorstores import FAISS from langchain.embeddings import OpenAIEmbeddings embeddings = OpenAIEmbeddings( openai_api_base="https://b-lost.com/v1", openai_api_key=os.getenv("BLOST_API_KEY") ) vector_store = FAISS.load_local("curriculum_db", embeddings) def retrieve_relevant_content(query, student_state): # Filter by student mastery level filters = {"difficulty": student_state["current_level"]} docs = vector_store.similarity_search(query, k=3, filter=filters) return docs The operational trap: RAG retrieval added 80-120ms latency, pushing total response time past 400ms. When benchmarking multi-provider fallback through the B-Lost relay, TTFT (time-to-first-token) dropped by 35ms with zero token truncation by caching embeddings client-side. More critical: retrieved content occasionally contradicted the knowledge state's difficulty estimates. A student flagged as struggling with algebra would receive calculus-level word problems because RAG pulled the highest-similarity docs without respecting mastery boundaries. We had to implement a two-stage filter: def filtered_retrieval(query, student_state): candidates = vector_store.similarity_search(query, k=10) # Post-filter by knowledge state compatibility valid_docs = [ doc for doc in candidates if doc.metadata["difficulty"] <= student_state["max_difficulty"] ] return valid_docs[:3] Debugging Knowledge Drift with CLI Probes Knowledge state vectors drift over time if interaction labels are noisy (students guessing correctly, or making careless errors on mastered concepts). We built a CLI diagnostic: #!/bin/bash # deeptutor-drift-probe.sh STUDENT_ID=$1 REDIS_KEY="deeptutor:student:${STUDENT_ID}:state" echo "Fetching state vector..." STATE=$(redis-cli GET $REDIS_KEY) echo "Analyzing interaction history..." HISTORY=$(redis-cli LRANGE "deeptutor:student:${STUDENT_ID}:history" 0 -1) python3 << EOF import json import numpy as np state = json.loads('$STATE') history = json.loads('$HISTORY') # Calculate concept mastery variance concept_scores = [] for interaction in history[-20:]: concept_scores.append(interaction['correctness']) variance = np.var(concept_scores) print(f"Mastery variance (last 20): {variance:.3f}") if variance > 0.25: print("WARNING: High variance detected - potential label noise") EOF Run every 50 interactions per student. Variance above 0.25 triggered manual review—often revealing mislabeled interactions or students rushing through questions. The Unresolved Question: Long-Term State Decay DeepTutor assumes lifelong learning, but our longest pilot ran 6 months. Student knowledge states grew to 500+ interaction records. At what point does historical data become noise rather than signal? We experimented with exponential decay weights: def calculate_weighted_state(history, decay_rate=0.95): weighted_sum = 0 total_weight = 0 for i, interaction in enumerate(reversed(history)): weight = decay_rate ** i weighted_sum += interaction['mastery_delta'] * weight total_weight += weight return weighted_sum / total_weight Decay rate tuning (0.90 vs 0.95 vs 0.98) changed student progression pacing by 15-20%. No consensus yet on optimal values—likely depends on subject domain and student age group. How do you balance honoring past struggles against recognizing genuine growth?

🚀 Your personal AI tutor, for free! DeepTutor helps you learn difficult topics through personalized explanations, interactive learning & step-by-step guidance. GitHub: hkuds/deeptutor

在 DeepTutor 中，输入一个主题，就能生成个性化的视频讲解。AI 的回答越来越长，但学习不一定要靠读大段文字。新版DeepTutor( HKUDS/DeepTutor ) 也是希望让个性化学习变得直观，更顺手：先通过动画理解内容，记笔记时遇到不明白的地方，选中那一段就能问 AI。讨论会留在原处，复习时随时回看，也能接着之前的问题继续学。

깃허브 인기 주식 AI 5개를 합쳤다｜클로드로 자동매매 만들기 깃허브에서 유명한 주식 AI 5개는 전부 해외 시장 기준입니다. 한국장 규칙(호가단위·±30% 가격제한·거래세·외국인·기관 수급·DART 공시·테마)에 맞게 바꿔서, AlphaForge 노드로 하나씩 연결해 실제로 돌려봤습니다. 📌 인프런 강의 「AI로 만드는 KR/US 퀀트 트레이딩 노드 분석 시스템 (바이브코딩)」 30% 할인 쿠폰 👉 inf.run (쿠폰 소진 시 만료) ⏱ 챕터 0:00 인트로 — 깃허브 주식 AI를 한국장으로 0:54 1장 TradingAgents — 강세·약세 토론과 승인 6:51 2장 Kronos — 5일 확률 예측과 계획표 10:52 3장 Vibe-Trading — 테마와 미래 정보 함정 14:02 4장 claude-trading-skills — 시장 폭·FTD·천장 위험 17:20 5장 day-trader-skill — 수급 연속 매수와 KRX 규칙 20:12 마무리 — 다 합치면, 그리고 자동매매 🧩 장별로 한 것 · 1장 TradingAgents — 수급·공시·차트·비중·강세/약세 토론·트레이더·리스크 승인을 노드 8개로 조립 (2,768 → 638 → 603 → 33 → 5) · 2장 Kronos — 5거래일 확률 예측 + KRX 호가·손익 계획표 (2,768 → 282 → 10) · 3장 Vibe-Trading — 테마로 고르기, '주도 테마'를 쓰면 미래 정보라 막히는 이유 · 4장 claude-trading-skills — 시장 폭 16.6(극약세) · 바닥 확인(FTD) · 천장 위험 85.5(위기) · 5장 day-trader-skill — 외국인·기관 연속 순매수(2,768 → 26)와 상한가·호가 규칙 · 마무리 — 추천 조합 한 번으로 2,768 → 141, 그리고 증권사 API 주문 노드로 자동매매까지 잇는 자리 📊 데이터 · 모든 숫자는 2026-10-29 종가 데이터로 직접 실행한 결과입니다. · 시세·종목 KRX 공식 / 수급·일봉 한국투자증권 Open API(조회 전용) / 공시 OpenDART / 테마 네이버 · 자동매매는 주문 노드를 붙여야 하며(아직 없음), 꼭 모의투자부터, 사람의 승인을 거쳐서 하세요. 🔓 영상에 쓴 노드·자료: [공개 링크] 🔗 원본 저장소 TradingAgents TauricResearch/TradingAgents Kronos shiyu-coder/Kronos Vibe-Trading HKUDS/Vibe-Trading claude-trading-skills tradermonty/claude-trading-skills day-trader-skill Jakeschincariol/day-trader-skill ⚠ 교육용 콘텐츠이며 투자 권유가 아닙니다. 투자 판단과 책임은 본인에게 있습니다. #주식AI #퀀트트레이딩 #바이브코딩 #클로드주식 #클로드 [en-US] I Combined the 5 Most Popular Stock AI on GitHub | Building Automated Trading with Claude The top 5 stock AIs on GitHub are all based on overseas markets. I modified them to fit the Korean market rules (bid increments, ±30% price limits, transaction taxes, foreign and institutional supply and demand, DART disclosures, and themes), connected them one by one using AlphaForge nodes, and ran them in practice. 📌 30% Discount Coupon for the Infrun Course "Building a KR/US Quant Trading Node Analysis System with AI (Vibe Coding)" 👉 inf.run (Expires when coupons are exhausted) ⏱ Chapters 0:00 Intro — GitHub Stock AI to the Korean Market 0:54 Chapter 1 TradingAgents — Bullish/Bearish Discussion and Approval 6:51 Chapter 2 Kronos — 5-Day Probability Prediction and Schedules 10:52 Chapter 3 Vibe-Trading — Themes and Future Information Pitfalls 14:02 Chapter 4 claude-trading-skills — Market Width, FTD, and Ceiling Risk 17:20 Chapter 5 day-trader-skill — Supply and Demand Continuous Buying and KRX Rules 20:12 Wrap-up — All in One, and Automated Trading 🧩 What We Did by Chapter · Chapter 1 TradingAgents — Assembling Supply/Demand, Disclosures, Charts, Weights, Bullish/Bearish Discussions, Traders, and Risk Approvals into 8 Nodes (2,768 → 638 → 603 → 33 → 5) · Chapter 2 Kronos — 5-Trading Day Probability Prediction + KRX Bid/Loss Plan (2,768 → 282 → 10) · Chapter 3 Vibe-Trading — Choosing by Theme, Why Using 'Leading Themes' Gets Blocked Due to Future Information · Chapter 4 claude-trading-skills — Market Width 16.6 (Extreme Bearish) · Bottom Confirmation (FTD) · Ceiling Risk 85.5 (Crisis) · Chapter 5 day-trader-skill — Consecutive Net Buying by Foreigners and Institutions (2,768 → 26) and Upper Limit/Bid Rules · Wrap-up — From 2,768 to 141 with a Single Recommended Combination, and Even Automated Trading via Brokerage API Order Nodes Connecting Points 📊 Data · All numbers are the results of direct execution using closing price data from October 29, 2026. · Market Prices/Stocks: KRX Formulas / Supply/Demand/Daily Charts: Korea Investment & Securities Open API (Inquiry Only) / Disclosures: OpenDART / Themes: Naver · Automated trading requires attaching an order node (not available yet). Please be sure to start with a simulated investment and obtain human approval. 🔓 Nodes and materials used in the video: [Public Links] 🔗 Original Repositories TradingAgents TauricResearch/TradingAgents Kronos shiyu-coder/Kronos Vibe-Trading HKUDS/Vibe-Trading claude-trading-skills tradermonty/claude-trading-skills day-trader-skill Jakeschincariol/day-trader-skill ⚠ This is educational content and not an investment recommendation. Investment decisions and responsibilities lie with the individual. #StockAI #QuantTrading #VibeCoding #ClaudeStock #Claude [id] Menggabungkan 5 AI Analisis Saham Terpopuler di GitHub | Membuat Trading Otomatis dengan Claude Lima AI saham teratas di GitHub semuanya berbasis pasar luar negeri. Saya memodifikasinya agar sesuai dengan aturan pasar Korea (kenaikan penawaran, batas harga ±30%, pajak transaksi, penawaran dan permintaan asing dan institusional, pengungkapan DART, dan tema), menghubungkannya satu per satu menggunakan node AlphaForge, dan menjalankannya dalam praktik. 📌 Kupon Diskon 30% untuk Kursus Infrun "Membangun Sistem Analisis Node Trading Kuantitatif KR/US dengan AI (Vibe Coding)" 👉 inf.run (Berakhir setelah kupon habis) ⏱ Bab 0:00 Pendahuluan — GitHub Stock AI untuk Pasar Korea 0:54 Bab 1 TradingAgents — Diskusi dan Persetujuan Bullish/Bearish 6:51 Bab 2 Kronos — Prediksi Probabilitas 5 Hari dan Jadwal 10:52 Bab 3 Vibe-Trading — Tema dan Jebakan Informasi Masa Depan 14:02 Bab 4 claude-trading-skills — Lebar Pasar, FTD, dan Risiko Batas Atas 17:20 Bab 5 day-trader-skill — Penawaran dan Permintaan Pembelian Berkelanjutan dan Aturan KRX 20:12 Ringkasan — Semua dalam Satu, dan Perdagangan Otomatis 🧩 Apa yang Kami Lakukan per Bab · Bab 1 TradingAgents — Menggabungkan Penawaran/Permintaan, Pengungkapan, Grafik, Bobot, Diskusi Bullish/Bearish, Trader, dan Persetujuan Risiko ke dalam 8 Node (2.768 → 638 → 603 → 33 → 5) · Bab 2 Kronos — Prediksi Probabilitas 5 Hari Perdagangan + Rencana Penawaran/Kerugian KRX (2.768 → 282 → 10) · Bab 3 Vibe-Trading — Memilih Berdasarkan Tema, Mengapa Penggunaan 'Tema Utama' Diblokir Karena Informasi Masa Depan · Bab 4 claude-trading-skills — Lebar Pasar 16,6 (Bearish Ekstrem) · Konfirmasi Bawah (FTD) · Risiko Batas Atas 85.5 (Krisis) · Bab 5 Keterampilan Daytrader — Pembelian Net Berurutan oleh Investor Asing dan Institusi (2.768 → 26) dan Batas Atas/Aturan Penawaran · Ringkasan — Dari 2.768 menjadi 141 dengan Satu Kombinasi Rekomendasi, dan Bahkan Perdagangan Otomatis melalui Node Pesanan API Broker yang Menghubungkan Titik 📊 Data · Semua angka adalah hasil eksekusi langsung menggunakan data harga penutupan dari 29 Oktober 2026. · Harga Pasar/Saham: Rumus KRX / Penawaran/Permintaan/Grafik Harian: API Terbuka Investasi & Sekuritas Korea (Hanya Permintaan) / Pengungkapan: OpenDART / Tema: Naver · Perdagangan otomatis memerlukan pemasangan node pesanan (belum tersedia). Pastikan untuk memulai dengan investasi simulasi dan mendapatkan persetujuan manusia. 🔓 Node dan material yang digunakan dalam video: [Tautan Publik] 🔗 Repositori Asli TradingAgents TauricResearch/TradingAgents Kronos shiyu-coder/Kronos Vibe-Trading HKUDS/Vibe-Trading claude-trading-skills tradermonty/claude-trading-skills day-trader-skill Jakeschincariol/day-trader-skill ⚠ Ini adalah konten edukatif dan bukan rekomendasi investasi. Keputusan dan tanggung jawab investasi berada di tangan individu. #StockAI #QuantTrading #VibeCoding #ClaudeStock #Claude

9 GitHub Repos Blowing Up Right Now (I Tested Every One) 📈 Agentic OS + Community: bit.ly 🩵 Free Skill: bit.ly 👋 Howdy Hey, I'm Jack, I built and sold my last tech startup with a Gazillion customers, now I'm building AI startups and I share stuff that works. I quit my corporate job to do my start-up and have never looked back since. 💬 Overview Nine GitHub repos that help you build better apps, create beautiful websites, find jobs and learn faster. I tested every one, and every link is below. Elephant-Goldfish helps Claude catch mistakes before it writes code. BAMA builds native Mac apps. Open UI turns AI answers into interactive charts and dashboards. claude-seo audits your website and gives you a plan to improve it. Postiz puts your social posts on one calendar. ai-job-search matches jobs to your CV and helps tailor your applications. DeepTutor turns a PDF into a tutor that explains the hard parts and quizzes you. Then I put Hallmark and MotionSites up against the same website brief to see which gives the best result. Plus, Honcho gives your AI memory across sessions, and Oxylabs gives your agents live web data for research. Links 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (code jack for $10 credit) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io ($5 free credit + 20% off with code JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Full Course & Community: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Stamps 0:00 - Intro 0:28 - Level 1: Elephant-Goldfish 2:52 - Bonus: Honcho 4:58 - Level 2: Open UI 6:20 - Level 3: BAMA 9:18 - Level 4: claude-seo 11:33 - Bonus: Oxylabs 12:58 - Level 5: Hallmark 14:48 - Level 6: MotionSites 15:58 - Level 7: Postiz 17:25 - Level 8: ai-job-search 18:47 - Level 9: DeepTutor 19:33 - Outro ▶️ Watch Next 7 Free Repos That Make Claude So Good It Feels Illegal: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [ar] 9 مستودعات GitHub تحقق انتشاراً واسعاً الآن (لقد جربت كلاً منها) 📈 نظام التشغيل Agentic + المجتمع: bit.ly 🩵 مهارة مجانية: bit.ly 👋 أهلًا وسهلًا أنا جاك، أسستُ شركتي التقنية الناشئة الأخيرة وبعتها، والتي حققت نجاحًا باهرًا مع عدد هائل من العملاء. الآن، أعمل على تأسيس شركات ناشئة في مجال الذكاء الاصطناعي، وأشارككم ما أثبت فعاليته. تركتُ وظيفتي في الشركة لأتفرغ لشركتي الناشئة، ولم أندم على ذلك أبدًا. 💬 نظرة عامة تسعة مستودعات على GitHub تُساعدك على بناء تطبيقات أفضل، وإنشاء مواقع ويب جذابة، وإيجاد وظائف، والتعلم بشكل أسرع. لقد اختبرتُ جميعها، وجميع الروابط موجودة أدناه. يُساعد Elephant-Goldfish برنامج Claude على اكتشاف الأخطاء قبل كتابة الكود. يُنشئ BAMA تطبيقات أصلية لنظام Mac. يُحوّل Open UI إجابات الذكاء الاصطناعي إلى رسوم بيانية ولوحات تحكم تفاعلية. يُدقّق claude-seo موقعك الإلكتروني ويُقدّم لك خطة لتحسينه. يجمع Postiz منشوراتك على مواقع التواصل الاجتماعي في تقويم واحد. يُطابق نظام البحث عن وظائف بالذكاء الاصطناعي الوظائف مع سيرتك الذاتية ويساعدك على تخصيص طلباتك. يحوّل DeepTutor ملف PDF إلى مُعلّم يشرح لك الأجزاء الصعبة ويختبرك. ثم قارنتُ بين Hallmark وMotionSites باستخدام نفس مواصفات الموقع الإلكتروني لمعرفة أيّهما يُقدّم أفضل النتائج. بالإضافة إلى ذلك، يُوفّر Honcho لذكائك الاصطناعي ذاكرة تخزين مؤقتة عبر الجلسات، ويُزوّد ​​Oxylabs وكلاءك ببيانات الويب المباشرة لأغراض البحث. روابط 🐘 نموذج سمكة الفيل الذهبية: google/elephant-goldfish-model 🧠 هونشو: go.honcho.dev (استخدم الكود jack للحصول على رصيد بقيمة 10 دولارات) 🍎 باما: daveebbelaar/bama 📊 واجهة المستخدم المفتوحة: thesysdev/openui 🔎 كلود-سيو: AgriciDaniel/claude-seo 🌐 أوكسي لابس: oxylabs.io (رصيد مجاني بقيمة 5 دولارات + خصم 20% باستخدام الكود JACK) 📅 بوستيز: gitroomhq/postiz-app 💼 البحث عن وظائف بالذكاء الاصطناعي: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + دورة كاملة ومجتمع: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai الفواصل الزمنية 0:00 - مقدمة 0:28 - المستوى 1: Elephant-Goldfish 2:52 - إضافي: Honcho 4:58 - المستوى 2: Open UI 6:20 - المستوى 3: BAMA 9:18 - المستوى 4: claude-seo 11:33 - إضافي: Oxylabs 12:58 - المستوى 5: هولمارك 14:48 - المستوى 6: مواقع الحركة 15:58 - المستوى 7: بوستيز 17:25 - المستوى 8: البحث عن وظائف بالذكاء الاصطناعي 18:47 - المستوى 9: ديب تيوتور 19:33 - الخاتمة ▶️ شاهد التالي 7 مستودعات مجانية تجعل كلود متميزًا لدرجة يصعب تصديقها: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [bn] এই মুহূর্তে সাড়া ফেলা ৯টি গিটহাব রিপো (আমি সবগুলোই পরীক্ষা করেছি) 📈 Agentic OS + কমিউনিটি: bit.ly 🩵 ফ্রি স্কিল: bit.ly 👋 হ্যালো হে, আমি জ্যাক। আমি আমার আগের টেক স্টার্টআপটি তৈরি করে বিক্রি করে দিয়েছি, যার অসংখ্য গ্রাহক ছিল। এখন আমি এআই (AI) স্টার্টআপ তৈরি করছি এবং কার্যকরী জিনিসগুলো শেয়ার করি। আমি আমার স্টার্টআপ শুরু করার জন্য কর্পোরেট চাকরি ছেড়ে দিয়েছি এবং তারপর থেকে আর পিছনে ফিরে তাকাইনি। 💬 সংক্ষিপ্ত বিবরণ নয়টি গিটহাব (GitHub) রিপোজিটরি যা আপনাকে আরও ভালো অ্যাপ তৈরি করতে, সুন্দর ওয়েবসাইট বানাতে, চাকরি খুঁজতে এবং দ্রুত শিখতে সাহায্য করে। আমি প্রতিটি পরীক্ষা করে দেখেছি এবং প্রতিটি লিঙ্ক নিচে দেওয়া আছে। Elephant-Goldfish, Claude-কে কোড লেখার আগেই ভুল ধরতে সাহায্য করে। BAMA নেটিভ ম্যাক (Mac) অ্যাপ তৈরি করে। Open UI, এআই-এর উত্তরগুলোকে ইন্টারেক্টিভ চার্ট এবং ড্যাশবোর্ডে পরিণত করে। claude-seo আপনার ওয়েবসাইট অডিট করে এবং এটি উন্নত করার জন্য একটি পরিকল্পনা দেয়। Postiz আপনার সোশ্যাল মিডিয়ার পোস্টগুলোকে একটি ক্যালেন্ডারে রাখে। ai-job-search আপনার সিভির সাথে চাকরি মিলিয়ে দেয় এবং আপনার আবেদনপত্রকে প্রয়োজন অনুযায়ী সাজাতে সাহায্য করে। DeepTutor একটি পিডিএফকে এমন একজন টিউটরে পরিণত করে, যে কঠিন অংশগুলো ব্যাখ্যা করে এবং আপনাকে কুইজ করায়। এরপর আমি একই ওয়েবসাইট ব্রিফের ভিত্তিতে Hallmark এবং MotionSites-কে তুলনা করে দেখি কোনটি সেরা ফলাফল দেয়। এছাড়াও, Honcho আপনার এআই-কে বিভিন্ন সেশন জুড়ে স্মৃতিশক্তি দেয়, এবং Oxylabs আপনার এজেন্টদের গবেষণার জন্য লাইভ ওয়েব ডেটা সরবরাহ করে। লিঙ্কসমূহ 🐘 এলিফ্যান্ট-গোল্ডফিশ: google/elephant-goldfish-model 🧠 হনচো: go.honcho.dev (জ্যাক কোড ব্যবহার করে $10 ক্রেডিট পান) 🍎 বামা: daveebbelaar/bama 📊 ওপেন ইউআই: thesysdev/openui 🔎 ক্লড-এসইও: AgriciDaniel/claude-seo 🌐 অক্সিল্যাবস: oxylabs.io (JACK কোড ব্যবহার করে $5 ফ্রি ক্রেডিট + 20% ছাড়) 📅 পোস্টিজ: gitroomhq/postiz-app 💼 এআই-জব-সার্চ: ai-job-search-dev/ai-job-search 🎓 ডিপটিউটর: HKUDS/DeepTutor 🎨 হলমার্ক: Nutlope/hallmark 🎬 মোশনসাইটস: Motionsites-AI/claude-skill 📈 এজেন্টিক ওএস + সম্পূর্ণ কোর্স ও কমিউনিটি: bit.ly 🗣️ গ্লাইডো: get.glaido.com 🤖 ক্লদ: claude.ai স্ট্যাম্প ০:০০ - ভূমিকা ০:২৮ - লেভেল ১: এলিফ্যান্ট-গোল্ডফিশ ২:৫২ - বোনাস: হনচো ৪:৫৮ - লেভেল ২: ওপেন ইউআই ৬:২০ - লেভেল ৩: বামা ৯:১৮ - লেভেল ৪: ক্লদ-এসইও ১১:৩৩ - বোনাস: অক্সিল্যাবস ১২:৫৮ - লেভেল ৫: হলমার্ক ১৪:৪৮ - লেভেল ৬: মোশনসাইটস ১৫:৫৮ - লেভেল ৭: পোস্টিজ ১৭:২৫ - লেভেল ৮: এআই-জব-সার্চ ১৮:৪৭ - লেভেল ৯: ডিপটিউটর ১৯:৩৩ - আউটরো ▶️ পরবর্তী ভিডিও দেখুন ৭টি ফ্রি রিপো যা ক্লডকে এতটাই দক্ষ করে তোলে যে তা বেআইনি মনে হয়: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [de-DE] 9 GitHub-Repos, die gerade durch die Decke gehen (ich habe alle getestet) 📈 Agentic OS + Community: bit.ly 🩵 Kostenloser Skill: bit.ly 👋 Hallo! Hey, ich bin Jack. Ich habe mein letztes Tech-Startup mit unzähligen Kunden aufgebaut und verkauft. Jetzt baue ich KI-Startups auf und teile bewährte Lösungen. Ich habe meinen Job in einem Konzern gekündigt, um mein Startup zu gründen, und es nie bereut. 💬 Übersicht Neun GitHub-Repos, die dir helfen, bessere Apps zu entwickeln, ansprechende Websites zu erstellen, Jobs zu finden und schneller zu lernen. Ich habe jedes einzelne getestet, und alle Links findest du unten. Elephant-Goldfish hilft Claude, Fehler zu erkennen, bevor er Code schreibt. BAMA erstellt native Mac-Apps. Open UI verwandelt KI-Antworten in interaktive Diagramme und Dashboards. claude-seo analysiert deine Website und erstellt einen Plan zur Optimierung. Postiz bündelt deine Social-Media-Posts in einem Kalender. ai-job-search gleicht Jobs mit deinem Lebenslauf ab und hilft dir, deine Bewerbungen zu personalisieren. DeepTutor verwandelt PDFs in Tutoren, die schwierige Abschnitte erklären und Ihr Wissen abfragen. Anschließend habe ich Hallmark und MotionSites anhand desselben Website-Briefings verglichen, um zu sehen, welches Tool das beste Ergebnis liefert. Zusätzlich stellt Honcho Ihrer KI ein sitzungsübergreifendes Gedächtnis zur Verfügung, und Oxylabs liefert Ihren Agenten Live-Webdaten für Recherchen. Links 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (Code: jack für 10 $ Guthaben) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (5 $ Guthaben + 20 % Rabatt mit Code: JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Komplettkurs & Community: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Stempel 0:00 – Einführung 0:28 – Level 1: Elefant-Goldfisch 2:52 – Bonus: Honcho 4:58 – Level 2: Open UI 6:20 – Level 3: BAMA 9:18 – Level 4: claude-seo 11:33 – Bonus: Oxylabs 12:58 – Level 5: Hallmark 14:48 – Level 6: MotionSites 15:58 – Level 7: Postiz 17:25 – Level 8: KI-Jobsuche 18:47 – Level 9: DeepTutor 19:33 – Abspann ▶️ Nächstes Video ansehen 7 kostenlose Repositories, die Claude so gut machen, dass es fast schon illegal ist: youtu.be #GitHub #ClaudeCode #OpenSource #KI-Tools #Claude [es-US] 9 repos de GitHub que están arrasando ahora mismo (los probé todos) 📈 Sistema Operativo Agenético + Comunidad: bit.ly 🩵 Habilidad Gratuita: bit.ly 👋 ¡Hola! Soy Jack. Fundé y vendí mi última startup tecnológica con muchísimos clientes. Ahora estoy creando startups de IA y comparto soluciones que funcionan. Dejé mi trabajo corporativo para dedicarme a mi startup y no me he arrepentido. 💬 Resumen Nueve repositorios de GitHub que te ayudan a crear mejores apps, sitios web atractivos, encontrar trabajo y aprender más rápido. Los he probado todos y los enlaces están abajo. Elephant-Goldfish ayuda a Claude a detectar errores antes de escribir código. BAMA crea apps nativas para Mac. Open UI convierte las respuestas de la IA en gráficos y paneles interactivos. claude-seo analiza tu sitio web y te ofrece un plan para mejorarlo. Postiz organiza tus publicaciones en redes sociales en un calendario. ai-job-search relaciona las ofertas de empleo con tu CV y ​​te ayuda a personalizar tus solicitudes. DeepTutor convierte un PDF en un tutor que explica las partes difíciles y te pone a prueba. Luego comparé Hallmark y MotionSites con el mismo briefing web para ver cuál ofrece el mejor resultado. Además, Honcho mantiene la memoria de tu IA entre sesiones, y Oxylabs proporciona a tus agentes datos web en tiempo real para la investigación. Enlaces 🐘 Modelo de pez dorado elefante: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (código jack para obtener $10 de crédito) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io ($5 de crédito gratis + 20% de descuento con el código JACK) 📅 Postiz: gitroomhq/postiz-app 💼 Búsqueda de empleo con IA: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Curso completo y comunidad: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Marcas 0:00 - Introducción 0:28 - Nivel 1: Elephant-Goldfish 2:52 - Bonus: Honcho 4:58 - Nivel 2: Open UI 6:20 - Nivel 3: BAMA 9:18 - Nivel 4: claude-seo 11:33 - Bonus: Oxylabs 12:58 - Nivel 5: Hallmark 14:48 - Nivel 6: MotionSites 15:58 - Nivel 7: Postiz 17:25 - Nivel 8: ai-job-search 18:47 - Nivel 9: DeepTutor 19:33 - Despedida ▶️ Ver siguiente 7 repositorios gratuitos que hacen que Claude sea tan bueno que parece ilegal: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [fr-FR] 9 dépôts GitHub qui font le buzz en ce moment (je les ai tous testés) 📈 Système d'exploitation Agentic + Communauté : bit.ly 🩵 Compétence gratuite : bit.ly 👋 Salut ! Je suis Jack. J'ai créé et revendu ma dernière startup tech avec un nombre impressionnant de clients. Aujourd'hui, je développe des startups en IA et je partage des solutions qui fonctionnent. J'ai quitté mon emploi en entreprise pour me lancer dans l'aventure entrepreneuriale et je ne l'ai jamais regretté. 💬 Présentation Neuf dépôts GitHub pour vous aider à créer de meilleures applications, de superbes sites web, à trouver un emploi et à apprendre plus vite. Je les ai tous testés et vous trouverez tous les liens ci-dessous. Elephant-Goldfish aide Claude à détecter les erreurs avant même l'écriture du code. BAMA crée des applications natives pour Mac. Open UI transforme les réponses de l'IA en graphiques et tableaux de bord interactifs. claude-seo audite votre site web et vous propose un plan d'amélioration. Postiz centralise vos publications sur les réseaux sociaux dans un seul calendrier. ai-job-search associe les offres d'emploi à votre CV et vous aide à personnaliser vos candidatures. DeepTutor transforme un PDF en un tuteur qui explique les points complexes et vous interroge. J'ai ensuite comparé Hallmark et MotionSites sur le même cahier des charges web afin de déterminer lequel offre les meilleurs résultats. De plus, Honcho permet à votre IA de mémoriser les informations entre les sessions, et Oxylabs fournit à vos agents des données web en temps réel pour la recherche. Liens 🐘 Elephant-Goldfish : google/elephant-goldfish-model 🧠 Honcho : go.honcho.dev (code JACK pour 10 $ de crédit) 🍎 BAMA : daveebbelaar/bama 📊 Open UI : thesysdev/openui 🔎 claude-seo : AgriciDaniel/claude-seo 🌐 Oxylabs : oxylabs.io (5 $ de crédit offerts + 20 % de réduction avec le code JACK) 📅 Postiz : gitroomhq/postiz-app 💼 ai-job-search : ai-job-search-dev/ai-job-search 🎓 DeepTutor : HKUDS/DeepTutor 🎨 Hallmark : Nutlope/hallmark 🎬 MotionSites : Motionsites-AI/claude-skill 📈 Agentic OS + Formation complète et communauté : bit.ly 🗣️ Glaido : get.glaido.com 🤖 Claude : claude.ai Stamps 0:00 - Intro 0:28 - Niveau 1 : Elephant-Goldfish 2:52 - Bonus : Honcho 4:58 - Niveau 2 : Open UI 6:20 - Niveau 3 : BAMA 9:18 - Niveau 4 : claude-seo 11:33 - Bonus : Oxylabs 12:58 - Niveau 5 : Hallmark 14:48 - Niveau 6 : MotionSites 15:58 - Niveau 7 : Postiz 17:25 - Niveau 8 : ai-job-search 18:47 - Niveau 9 : DeepTutor 19:33 - Conclusion ▶️ À voir ensuite 7 dépôts gratuits qui rendent Claude tellement performant que c'en est presque illégal : youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [hi] इस समय धूम मचा रहे 9 GitHub रिपॉज़िटरी (मैंने हर एक को टेस्ट किया है) 📈 एजेंटिक ओएस + कम्युनिटी: bit.ly 🩵 मुफ़्त स्किल: bit.ly 👋 नमस्कार! मैं जैक हूँ। मैंने अपना पिछला टेक स्टार्टअप बनाया और बेचा, जिसके अनगिनत ग्राहक थे। अब मैं AI स्टार्टअप बना रहा हूँ और काम की चीज़ें शेयर करता हूँ। मैंने अपना स्टार्टअप शुरू करने के लिए कॉर्पोरेट नौकरी छोड़ दी और तब से पीछे मुड़कर नहीं देखा। 💬 संक्षिप्त जानकारी नौ GitHub रिपॉज़िटरी जो आपको बेहतर ऐप्स बनाने, खूबसूरत वेबसाइट बनाने, नौकरी ढूंढने और तेज़ी से सीखने में मदद करती हैं। मैंने हर एक का परीक्षण किया है, और हर लिंक नीचे दिया गया है। एलिफेंट-गोल्डफ़िश क्लाउड को कोड लिखने से पहले गलतियाँ पकड़ने में मदद करता है। BAMA नेटिव मैक ऐप्स बनाता है। ओपन यूआई AI के जवाबों को इंटरैक्टिव चार्ट और डैशबोर्ड में बदलता है। claude-seo आपकी वेबसाइट का ऑडिट करता है और उसे बेहतर बनाने के लिए आपको एक प्लान देता है। पोस्टिज़ आपके सोशल मीडिया पोस्ट को एक कैलेंडर पर दिखाता है। एआई जॉब सर्च आपके सीवी से नौकरियों का मिलान करता है और आपके आवेदन को अनुकूलित करने में मदद करता है। डीप ट्यूटर एक पीडीएफ को ट्यूटर में बदल देता है जो कठिन भागों को समझाता है और आपसे प्रश्न पूछता है। फिर मैंने हॉलमार्क और मोशनसाइट्स को एक ही वेबसाइट ब्रीफ के आधार पर परखा ताकि यह पता चल सके कि कौन सा बेहतर परिणाम देता है। इसके अलावा, होंचो आपके एआई को सत्रों के दौरान मेमोरी प्रदान करता है, और ऑक्सिलैब्स आपके एजेंटों को शोध के लिए लाइव वेब डेटा प्रदान करता है। लिंक 🐘 एलिफेंट-गोल्डफिश: google/elephant-goldfish-model 🧠 होंचो: go.honcho.dev (कोड jack का उपयोग करके $10 का क्रेडिट प्राप्त करें) 🍎 बामा: daveebbelaar/bama 📊 ओपन यूआई: thesysdev/openui 🔎 क्लाउड-एसईओ: AgriciDaniel/claude-seo 🌐 ऑक्सीलैब्स: oxylabs.io (कोड JACK के साथ $5 का मुफ़्त क्रेडिट + 20% की छूट) 📅 पोस्टिज़: gitroomhq/postiz-app 💼 एआई-जॉब-सर्च: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + पूर्ण पाठ्यक्रम और समुदाय: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai स्टैम्प 0:00 - परिचय 0:28 - स्तर 1: Elephant-Goldfish 2:52 - बोनस: Honcho 4:58 - स्तर 2: Open UI 6:20 - स्तर 3: BAMA 9:18 - स्तर 4: claude-seo 11:33 - बोनस: Oxylabs 12:58 - स्तर 5: हॉलमार्क 14:48 - लेवल 6: मोशनसाइट्स 15:58 - लेवल 7: पोस्टिज़ 17:25 - लेवल 8: एआई-जॉब-सर्च 18:47 - लेवल 9: डीपट्यूटर 19:33 - आउट्रो ▶️ अगला वीडियो देखें 7 मुफ़्त रिपॉज़िटरी जो क्लाउड को इतना बेहतरीन बनाती हैं कि यह गैरकानूनी लगता है: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [id] 9 Repositori GitHub yang Sedang Viral Saat Ini (Saya Menguji Semuanya) 📈 Agentic OS + Komunitas: bit.ly 🩵 Skill Gratis: bit.ly 👋 Halo Hai, saya Jack, saya membangun dan menjual startup teknologi terakhir saya dengan jutaan pelanggan, sekarang saya membangun startup AI dan saya berbagi hal-hal yang berhasil. Saya berhenti dari pekerjaan korporat saya untuk menjalankan startup saya dan tidak pernah menyesalinya sejak itu. 💬 Gambaran Umum Sembilan repositori GitHub yang membantu Anda membangun aplikasi yang lebih baik, membuat situs web yang indah, menemukan pekerjaan, dan belajar lebih cepat. Saya telah menguji semuanya, dan setiap tautan ada di bawah ini. Elephant-Goldfish membantu Claude mendeteksi kesalahan sebelum menulis kode. BAMA membangun aplikasi Mac asli. Open UI mengubah jawaban AI menjadi grafik dan dasbor interaktif. claude-seo mengaudit situs web Anda dan memberi Anda rencana untuk memperbaikinya. Postiz menempatkan postingan media sosial Anda dalam satu kalender. ai-job-search mencocokkan pekerjaan dengan CV Anda dan membantu menyesuaikan lamaran Anda. DeepTutor mengubah PDF menjadi tutor yang menjelaskan bagian-bagian sulit dan memberikan kuis kepada Anda. Kemudian saya membandingkan Hallmark dan MotionSites dengan brief situs web yang sama untuk melihat mana yang memberikan hasil terbaik. Selain itu, Honcho memberikan memori AI Anda di seluruh sesi, dan Oxylabs memberikan data web langsung kepada agen Anda untuk riset. Tautan 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (kode jack untuk kredit $10) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (kredit gratis $5 + diskon 20% dengan kode JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Kursus Lengkap & Komunitas: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Stempel 0:00 - Pendahuluan 0:28 - Level 1: Elephant-Goldfish 2:52 - Bonus: Honcho 4:58 - Level 2: Open UI 6:20 - Level 3: BAMA 9:18 - Level 4: claude-seo 11:33 - Bonus: Oxylabs 12:58 - Level 5: Hallmark 14:48 - Level 6: MotionSites 15:58 - Level 7: Postiz 17:25 - Level 8: ai-job-search 18:47 - Level 9: DeepTutor 19:33 - Penutup ▶️ Tonton Selanjutnya 7 Repositori Gratis yang Membuat Claude Begitu Bagus Hingga Terasa Ilegal: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [it] 9 repository GitHub che stanno spopolando ora (li ho testati tutti) 📈 Agentic OS + Community: bit.ly 🩵 Skill gratuita: bit.ly 👋 Ciao Ciao, sono Jack, ho creato e venduto la mia ultima startup tecnologica con un'infinità di clienti, ora sto creando startup di intelligenza artificiale e condivido cose che funzionano. Ho lasciato il mio lavoro in azienda per dedicarmi alla mia startup e non me ne sono mai pentito. 💬 Panoramica Nove repository GitHub che ti aiutano a creare app migliori, siti web accattivanti, trovare lavoro e imparare più velocemente. Li ho testati tutti e i link sono qui sotto. Elephant-Goldfish aiuta Claude a individuare gli errori prima di scrivere il codice. BAMA crea app native per Mac. Open UI trasforma le risposte dell'IA in grafici e dashboard interattivi. claude-seo analizza il tuo sito web e ti fornisce un piano per migliorarlo. Postiz organizza i tuoi post sui social media in un unico calendario. AI-job-search abbina le offerte di lavoro al tuo CV e ti aiuta a personalizzare le candidature. DeepTutor trasforma un PDF in un tutor che spiega le parti più difficili e ti mette alla prova con dei quiz. Ho poi messo a confronto Hallmark e MotionSites con lo stesso brief per il sito web, per vedere quale offre i risultati migliori. Inoltre, Honcho mantiene la memoria dell'IA tra una sessione e l'altra, e Oxylabs fornisce ai tuoi agenti dati web in tempo reale per le ricerche. Link 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (codice jack per 10$ di credito) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (5$ di credito gratuito + 20% di sconto con il codice JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Corso completo e community: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Timps 0:00 - Introduzione 0:28 - Livello 1: Elephant-Goldfish 2:52 - Bonus: Honcho 4:58 - Livello 2: Open UI 6:20 - Livello 3: BAMA 9:18 - Livello 4: claude-seo 11:33 - Bonus: Oxylabs 12:58 - Livello 5: Hallmark 14:48 - Livello 6: MotionSites 15:58 - Livello 7: Postiz 17:25 - Livello 8: ai-job-search 18:47 - Livello 9: DeepTutor 19:33 - Conclusione ▶️ Guarda il prossimo video 7 repository gratuiti che rendono Claude così valido da sembrare illegale: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [iw] 9 מאגרי GitHub ששורפים את הרשת כרגע (בדקתי את כולם) 📈 Agentic OS + קהילה: bit.ly 🩵 מיומנות חינמית: bit.ly 👋 שלום היי, אני ג'ק, בניתי ומכרתי את הסטארט-אפ הטכנולוגי האחרון שלי עם מיליון לקוחות, עכשיו אני בונה סטארט-אפים של בינה מלאכותית ואני משתף דברים שעובדים. עזבתי את עבודתי התאגידית כדי לעשות את הסטארט-אפ שלי ומאז לא הסתכלתי לאחור. 💬 סקירה כללית תשעה מאגרי GitHub שעוזרים לך לבנות אפליקציות טובות יותר, ליצור אתרים יפים, למצוא עבודות וללמוד מהר יותר. בדקתי כל אחד מהם, וכל קישור נמצא למטה. Elephant-Goldfish עוזר ל-Claude לזהות טעויות לפני שהוא כותב קוד. BAMA בונה אפליקציות Mac מקוריות. ממשק משתמש פתוח הופך תשובות בינה מלאכותית לתרשימים ולוחות מחוונים אינטראקטיביים. claude-seo מבקר את האתר שלך ונותן לך תוכנית לשיפורו. Postiz שם את הפוסטים החברתיים שלך בלוח שנה אחד. ai-job-search מתאים משרות לקורות החיים שלך ועוזר להתאים אישית את הבקשות שלך. DeepTutor הופך קובץ PDF למורה פרטי שמסביר את החלקים הקשים ומבחן אותך. לאחר מכן אני מניח את Hallmark ו-MotionSites מול אותו מדריך אתר כדי לראות איזה מהם נותן את התוצאה הטובה ביותר. בנוסף, Honcho מספק את זיכרון הבינה המלאכותית שלך לאורך כל הסשנים, ו-Oxylabs מספק לסוכנים שלך נתוני אינטרנט בזמן אמת למחקר. קישורים 🐘 פיל-דג זהב: google/elephant-goldfish-model 🧠 הונצ'ו: go.honcho.dev (קוד ג'ק לקבלת זיכוי של 10 דולר) 🍎 באמה: daveebbelaar/bama 📊 ממשק משתמש פתוח: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 אוקסילבס: oxylabs.io (זיכוי חינם של 5 דולר + 20% הנחה עם הקוד JACK) 📅 פוסטיז: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 דיפ-טוטור: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 מערכת הפעלה Agentic + קורס מלא וקהילה: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai חותמות 0:00 - מבוא 0:28 - רמה 1: פיל-דג זהב 2:52 - בונוס: Honcho 4:58 - רמה 2: ממשק משתמש פתוח 6:20 - רמה 3: BAMA 9:18 - רמה 4: claude-seo 11:33 - בונוס: Oxylabs 12:58 - רמה 5: Hallmark 14:48 - רמה 6: MotionSites 15:58 - רמה 7: Postiz 17:25 - רמה 8: חיפוש-עבודה-בבינה מלאכותית 18:47 - רמה 9: DeepTutor 19:33 - סיום ▶️ צפו בהמשך 7 מאגרי מידע חינמיים שהופכים את קלוד לכל כך טוב שזה מרגיש לא חוקי: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [ja] 今話題のGitHubリポジトリ9選（全件検証済み） 📈 Agentic OS + コミュニティ: bit.ly 🩵 無料スキル: bit.ly 👋 こんにちは ジャックです。以前、数百万もの顧客を抱えるテクノロジー系スタートアップを立ち上げ、売却しました。現在はAIスタートアップを立ち上げながら、実際に役立つ情報を共有しています。会社員生活を辞めてスタートアップを始めて以来、後悔したことは一度もありません。 💬 概要 より良いアプリ開発、美しいウェブサイト制作、仕事探し、そして学習スピードアップに役立つ9つのGitHubリポジトリをご紹介します。すべて実際に試用済みで、リンクは下記に掲載しています。 Elephant-Goldfishは、Claudeがコードを書く前に間違いを検出するのに役立ちます。BAMAはネイティブMacアプリを開発します。Open UIはAIの回答をインタラクティブなグラフやダッシュボードに変換します。claude-seoはウェブサイトを監査し、改善プランを提供します。 Postizはソーシャルメディアの投稿をカレンダーにまとめて管理します。 ai-job-searchはあなたの履歴書に基づいて求人情報をマッチングし、応募書類の作成を支援します。DeepTutorはPDFファイルをチューターに変換し、難しい部分を解説し、クイズ形式で理解度を確認させてくれます。 次に、HallmarkとMotionSitesを同じウェブサイト作成要件に適用し、どちらがより良い結果をもたらすかを検証しました。 さらに、HonchoはAIにセッション間の記憶機能を提供し、Oxylabsはエージェントにリアルタイムのウェブデータを提供して調査を支援します。 リンク 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (コード「jack」で10ドルのクレジットを獲得) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (コード「JACK」で5ドルの無料クレジットと20%オフ) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + フルコース＆コミュニティ: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai スタンプ 0:00 - はじめに 0:28 - レベル1: ゾウと金魚 2:52 - ボーナス: Honcho 4:58 - レベル2: Open UI 6:20 - レベル3: BAMA 9:18 - レベル4: claude-seo 11:33 - ボーナス: Oxylabs 12:58 - レベル5: Hallmark 14:48 - レベル6：MotionSites 15:58 - レベル7：Postiz 17:25 - レベル8：ai-job-search 18:47 - レベル9：DeepTutor 19:33 - エンディング ▶️ 次の動画を見る Claudeを驚くほど便利にする7つの無料リポジトリ：youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [ko] 지금 가장 뜨거운 GitHub 저장소 9곳 (전부 직접 테스트해 봄) 📈 Agentic OS + 커뮤니티: bit.ly 🩵 무료 스킬: bit.ly 👋 안녕하세요! 저는 Jack입니다. 이전에 수많은 고객을 확보한 기술 스타트업을 설립하고 매각했으며, 현재는 AI 스타트업을 운영하며 효과적인 방법들을 공유하고 있습니다. 회사 생활을 그만두고 스타트업을 시작한 이후로 한 번도 후회한 적이 없습니다. 💬 개요 앱 개발, 웹사이트 제작, 구직 활동, 학습 속도 향상에 도움이 되는 9개의 GitHub 저장소를 소개합니다. 모든 저장소를 직접 테스트했으며, 링크는 아래에 있습니다. Elephant-Goldfish는 Claude가 코드를 작성하기 전에 오류를 잡아낼 수 있도록 도와줍니다. BAMA는 네이티브 Mac 앱을 개발합니다. Open UI는 AI 답변을 인터랙티브 차트와 대시보드로 변환합니다. claude-seo는 웹사이트를 분석하고 개선 방안을 제시합니다. Postiz는 소셜 미디어 게시물을 하나의 캘린더에 정리해 줍니다. AI 기반 구직 검색은 사용자의 이력서에 맞는 일자리를 찾아 맞춤형 지원서를 작성하도록 도와줍니다. DeepTutor는 PDF 파일을 분석하여 어려운 부분을 설명하고 퀴즈를 내는 튜터 역할을 합니다. 그런 다음 Hallmark와 MotionSites를 동일한 웹사이트 구축 과제에 투입하여 어떤 프로그램이 더 나은 결과를 보여주는지 비교해 보았습니다. 또한 Honcho는 AI에 세션 간 메모리 기능을 제공하고, Oxylabs는 에이전트에게 실시간 웹 데이터를 제공하여 조사 작업을 지원합니다. 링크 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (코드 'jack' 입력 시 $10 크레딧 제공) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (코드 'JACK' 입력 시 $5 무료 크레딧 + 20% 할인) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + 전체 강좌 및 커뮤니티: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai 스탬프 0:00 - 소개 0:28 - 레벨 1: Elephant-Goldfish 2:52 - 보너스: Honcho 4:58 - 레벨 2: Open UI 6:20 - 레벨 3: BAMA 9:18 - 레벨 4: claude-seo 11:33 - 보너스: Oxylabs 12:58 - 레벨 5: Hallmark 14:48 - 레벨 6: MotionSites 15:58 - 레벨 7: Postiz 17:25 - 레벨 8: ai-job-search 18:47 - 레벨 9: DeepTutor 19:33 - 마무리 ▶️ 다음 영상 보기 Claude를 너무 좋게 만들어주는 무료 저장소 7가지: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [ml] ഇപ്പോൾ തരംഗമാകുന്ന 9 GitHub റിപ്പോസിറ്ററികൾ (ഞാൻ എല്ലാം പരീക്ഷിച്ച് നോക്കി) 📈 ഏജന്റ് OS + കമ്മ്യൂണിറ്റി: bit.ly 🩵 സൗജന്യ വൈദഗ്ദ്ധ്യം: bit.ly 👋 ഹലോ ഹേയ്, ഞാൻ ജാക്ക് ആണ്, ഒരു ഗാസിലിയൻ ഉപഭോക്താക്കളുമായി ഞാൻ എന്റെ അവസാന ടെക് സ്റ്റാർട്ടപ്പ് നിർമ്മിച്ച് വിറ്റു, ഇപ്പോൾ ഞാൻ AI സ്റ്റാർട്ടപ്പുകൾ നിർമ്മിക്കുകയാണ്, പ്രവർത്തിക്കുന്ന കാര്യങ്ങൾ ഞാൻ പങ്കിടുന്നു. എന്റെ സ്റ്റാർട്ടപ്പ് ചെയ്യാൻ ഞാൻ എന്റെ കോർപ്പറേറ്റ് ജോലി ഉപേക്ഷിച്ചു, അതിനുശേഷം ഒരിക്കലും തിരിഞ്ഞുനോക്കിയിട്ടില്ല. 💬 അവലോകനം മികച്ച ആപ്പുകൾ നിർമ്മിക്കാനും മനോഹരമായ വെബ്‌സൈറ്റുകൾ സൃഷ്ടിക്കാനും ജോലികൾ കണ്ടെത്താനും വേഗത്തിൽ പഠിക്കാനും നിങ്ങളെ സഹായിക്കുന്ന ഒമ്പത് GitHub ശേഖരങ്ങൾ. ഞാൻ ഓരോന്നും പരീക്ഷിച്ചു, എല്ലാ ലിങ്കുകളും ചുവടെയുണ്ട്. കോഡ് എഴുതുന്നതിനുമുമ്പ് തെറ്റുകൾ കണ്ടെത്താൻ എലിഫന്റ്-ഗോൾഡ് ഫിഷ് ക്ലോഡിനെ സഹായിക്കുന്നു. BAMA നേറ്റീവ് മാക് ആപ്പുകൾ നിർമ്മിക്കുന്നു. ഓപ്പൺ UI AI ഉത്തരങ്ങളെ ഇന്ററാക്ടീവ് ചാർട്ടുകളായും ഡാഷ്‌ബോർഡുകളായും മാറ്റുന്നു. claude-seo നിങ്ങളുടെ വെബ്‌സൈറ്റ് ഓഡിറ്റ് ചെയ്യുകയും അത് മെച്ചപ്പെടുത്തുന്നതിനുള്ള ഒരു പദ്ധതി നൽകുകയും ചെയ്യുന്നു. Postiz നിങ്ങളുടെ സോഷ്യൽ പോസ്റ്റുകൾ ഒരു കലണ്ടറിൽ ഇടുന്നു. AI-job-search നിങ്ങളുടെ സിവിയുമായി ജോലികൾ പൊരുത്തപ്പെടുത്തുകയും നിങ്ങളുടെ ആപ്ലിക്കേഷനുകൾ ക്രമീകരിക്കാൻ സഹായിക്കുകയും ചെയ്യുന്നു. DeepTutor ഒരു PDF-നെ ബുദ്ധിമുട്ടുള്ള ഭാഗങ്ങൾ വിശദീകരിക്കുന്ന ഒരു ട്യൂട്ടറായി മാറ്റുകയും നിങ്ങളെ ചോദ്യം ചെയ്യുകയും ചെയ്യുന്നു. പിന്നെ ഏതാണ് മികച്ച ഫലം നൽകുന്നതെന്ന് കാണാൻ ഞാൻ അതേ വെബ്‌സൈറ്റ് ബ്രീഫുമായി ഹാൾമാർക്കും മോഷൻസൈറ്റുകളും ചേർത്തു. കൂടാതെ, ഹോഞ്ചോ സെഷനുകളിലുടനീളം നിങ്ങളുടെ AI മെമ്മറി നൽകുന്നു, കൂടാതെ Oxylabs നിങ്ങളുടെ ഏജന്റുമാർക്ക് ഗവേഷണത്തിനായി തത്സമയ വെബ് ഡാറ്റ നൽകുന്നു. ലിങ്കുകൾ 🐘 എലിഫന്റ്-ഗോൾഡ്ഫിഷ്: google/elephant-goldfish-model 🧠 ഹോഞ്ചോ: go.honcho.dev ($10 ക്രെഡിറ്റിന് കോഡ് ജാക്ക്) 🍎 BAMA: daveebbelaar/bama 📊 ഓപ്പൺ UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io ($5 സൗജന്യ ക്രെഡിറ്റ് + JACK എന്ന കോഡിനൊപ്പം 20% കിഴിവ്) 📅 പോസ്റ്റിസ്: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 ഹാൾമാർക്ക്: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 ഏജന്റ് ഒഎസ് + പൂർണ്ണ കോഴ്‌സും കമ്മ്യൂണിറ്റിയും: bit.ly 🗣️ ഗ്ലൈഡോ: get.glaido.com 🤖 ക്ലോഡ്: claude.ai സ്റ്റാമ്പുകൾ 0:00 - ആമുഖം 0:28 - ലെവൽ 1: എലിഫന്റ്-ഗോൾഡ് ഫിഷ് 2:52 - ബോണസ്: ഹോഞ്ചോ 4:58 - ലെവൽ 2: ഓപ്പൺ യുഐ 6:20 - ലെവൽ 3: BAMA 9:18 - ലെവൽ 4: ക്ലോഡ്-സിഇഒ 11:33 - ബോണസ്: ഓക്‌സിലാബ്‌സ് 12:58 - ലെവൽ 5: ഹാൾമാർക്ക് 14:48 - ലെവൽ 6: MotionSites 15:58 - ലെവൽ 7: Postiz 17:25 - ലെവൽ 8: ai-job-search 18:47 - ലെവൽ 9: DeepTutor 19:33 - Outro ▶️ അടുത്തത് കാണുക ക്ലോഡിനെ വളരെ നല്ലതാക്കുന്ന 7 സൗജന്യ റിപ്പോകൾ നിയമവിരുദ്ധമാണെന്ന് തോന്നുന്നു: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [nl-NL] 9 GitHub-repo's die nu enorm populair zijn (Ik heb ze allemaal getest) 📈 Agentic OS + Community: bit.ly 🩵 Gratis Skill: bit.ly 👋 Hallo Hey, ik ben Jack. Ik heb mijn vorige tech-startup met een enorm klantenbestand opgebouwd en verkocht. Nu bouw ik AI-startups en deel ik de dingen die werken. Ik heb mijn baan bij een groot bedrijf opgezegd om mijn startup te beginnen en heb daar geen moment spijt van gehad. 💬 Overzicht Negen GitHub-repositories die je helpen betere apps te bouwen, prachtige websites te maken, banen te vinden en sneller te leren. Ik heb ze allemaal getest en alle links staan ​​hieronder. Elephant-Goldfish helpt Claude fouten te vinden voordat het code schrijft. BAMA bouwt native Mac-apps. Open UI zet AI-antwoorden om in interactieve grafieken en dashboards. claude-seo controleert je website en geeft je een plan om deze te verbeteren. Postiz zet je social media-berichten op één kalender. ai-job-search koppelt vacatures aan je cv en helpt je sollicitaties te personaliseren. DeepTutor maakt van een PDF een tutor die de lastige onderdelen uitlegt en je overhoort. Vervolgens heb ik Hallmark en MotionSites met dezelfde websiteopdracht vergeleken om te zien welke het beste resultaat oplevert. Daarnaast geeft Honcho je AI geheugen over meerdere sessies heen, en Oxylabs biedt je agents realtime webdata voor onderzoek. Links 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (code jack voor $10 tegoed) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io ($5 gratis tegoed + 20% korting met code JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Volledige cursus & community: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Stempels 0:00 - Intro 0:28 - Niveau 1: Olifant-Goudvis 2:52 - Bonus: Honcho 4:58 - Niveau 2: Open UI 6:20 - Niveau 3: BAMA 9:18 - Niveau 4: claude-seo 11:33 - Bonus: Oxylabs 12:58 - Niveau 5: Hallmark 14:48 - Niveau 6: MotionSites 15:58 - Niveau 7: Postiz 17:25 - Niveau 8: ai-job-search 18:47 - Niveau 9: DeepTutor 19:33 - Outro ▶️ Bekijk volgende video 7 gratis repositories die Claude zo goed maken dat het bijna illegaal aanvoelt: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [pa] 9 GitHub ਰੈਪੋ ਜੋ ਇਸ ਵੇਲੇ ਬਹੁਤ ਮਸ਼ਹੂਰ ਹੋ ਰਹੀਆਂ ਹਨ (ਮੈਂ ਹਰ ਇੱਕ ਨੂੰ ਟੈਸਟ ਕੀਤਾ) 📈 ਏਜੰਟਿਕ ਓਐਸ + ਕਮਿਊਨਿਟੀ: bit.ly 🩵 ਮੁਫ਼ਤ ਹੁਨਰ: bit.ly 👋 ਹਾਏ ਹਾਏ ਹੇ, ਮੈਂ ਜੈਕ ਹਾਂ, ਮੈਂ ਆਪਣਾ ਆਖਰੀ ਤਕਨੀਕੀ ਸਟਾਰਟਅੱਪ ਇੱਕ ਗੈਜ਼ੀਲੀਅਨ ਗਾਹਕਾਂ ਨਾਲ ਬਣਾਇਆ ਅਤੇ ਵੇਚਿਆ, ਹੁਣ ਮੈਂ ਏਆਈ ਸਟਾਰਟਅੱਪ ਬਣਾ ਰਿਹਾ ਹਾਂ ਅਤੇ ਮੈਂ ਕੰਮ ਕਰਨ ਵਾਲੀਆਂ ਚੀਜ਼ਾਂ ਸਾਂਝੀਆਂ ਕਰਦਾ ਹਾਂ। ਮੈਂ ਆਪਣਾ ਸਟਾਰਟ-ਅੱਪ ਕਰਨ ਲਈ ਆਪਣੀ ਕਾਰਪੋਰੇਟ ਨੌਕਰੀ ਛੱਡ ਦਿੱਤੀ ਅਤੇ ਉਦੋਂ ਤੋਂ ਕਦੇ ਪਿੱਛੇ ਮੁੜ ਕੇ ਨਹੀਂ ਦੇਖਿਆ। 💬 ਸੰਖੇਪ ਜਾਣਕਾਰੀ ਨੌਂ GitHub ਰਿਪੋ ਜੋ ਤੁਹਾਨੂੰ ਬਿਹਤਰ ਐਪਸ ਬਣਾਉਣ, ਸੁੰਦਰ ਵੈੱਬਸਾਈਟਾਂ ਬਣਾਉਣ, ਨੌਕਰੀਆਂ ਲੱਭਣ ਅਤੇ ਤੇਜ਼ੀ ਨਾਲ ਸਿੱਖਣ ਵਿੱਚ ਮਦਦ ਕਰਦੇ ਹਨ। ਮੈਂ ਹਰ ਇੱਕ ਦੀ ਜਾਂਚ ਕੀਤੀ, ਅਤੇ ਹਰ ਲਿੰਕ ਹੇਠਾਂ ਹੈ। ਐਲੀਫੈਂਟ-ਗੋਲਡਫਿਸ਼ ਕੋਡ ਲਿਖਣ ਤੋਂ ਪਹਿਲਾਂ ਕਲੌਡ ਨੂੰ ਗਲਤੀਆਂ ਫੜਨ ਵਿੱਚ ਮਦਦ ਕਰਦਾ ਹੈ। BAMA ਨੇਟਿਵ ਮੈਕ ਐਪਸ ਬਣਾਉਂਦਾ ਹੈ। ਓਪਨ UI AI ਜਵਾਬਾਂ ਨੂੰ ਇੰਟਰਐਕਟਿਵ ਚਾਰਟ ਅਤੇ ਡੈਸ਼ਬੋਰਡ ਵਿੱਚ ਬਦਲਦਾ ਹੈ। ਕਲਾਉਡ-ਸੀਓ ਤੁਹਾਡੀ ਵੈੱਬਸਾਈਟ ਦਾ ਆਡਿਟ ਕਰਦਾ ਹੈ ਅਤੇ ਤੁਹਾਨੂੰ ਇਸਨੂੰ ਬਿਹਤਰ ਬਣਾਉਣ ਦੀ ਯੋਜਨਾ ਦਿੰਦਾ ਹੈ। ਪੋਸਟਿਜ਼ ਤੁਹਾਡੀਆਂ ਸੋਸ਼ਲ ਪੋਸਟਾਂ ਨੂੰ ਇੱਕ ਕੈਲੰਡਰ 'ਤੇ ਰੱਖਦਾ ਹੈ। ਏਆਈ-ਜੌਬ-ਸਰਚ ਨੌਕਰੀਆਂ ਨੂੰ ਤੁਹਾਡੇ ਸੀਵੀ ਨਾਲ ਮੇਲ ਖਾਂਦਾ ਹੈ ਅਤੇ ਤੁਹਾਡੀਆਂ ਐਪਲੀਕੇਸ਼ਨਾਂ ਨੂੰ ਅਨੁਕੂਲ ਬਣਾਉਣ ਵਿੱਚ ਮਦਦ ਕਰਦਾ ਹੈ। ਡੀਪ ਟਿਊਟਰ ਇੱਕ PDF ਨੂੰ ਇੱਕ ਟਿਊਟਰ ਵਿੱਚ ਬਦਲਦਾ ਹੈ ਜੋ ਔਖੇ ਹਿੱਸਿਆਂ ਨੂੰ ਸਮਝਾਉਂਦਾ ਹੈ ਅਤੇ ਤੁਹਾਨੂੰ ਪੁੱਛਗਿੱਛ ਕਰਦਾ ਹੈ। ਫਿਰ ਮੈਂ ਹਾਲਮਾਰਕ ਅਤੇ ਮੋਸ਼ਨ ਸਾਈਟਸ ਨੂੰ ਉਸੇ ਵੈੱਬਸਾਈਟ ਸੰਖੇਪ ਦੇ ਵਿਰੁੱਧ ਰੱਖਦਾ ਹਾਂ ਤਾਂ ਜੋ ਇਹ ਦੇਖਿਆ ਜਾ ਸਕੇ ਕਿ ਕਿਹੜਾ ਸਭ ਤੋਂ ਵਧੀਆ ਨਤੀਜਾ ਦਿੰਦਾ ਹੈ। ਇਸ ਤੋਂ ਇਲਾਵਾ, ਹੋਂਚੋ ਸੈਸ਼ਨਾਂ ਵਿੱਚ ਤੁਹਾਡੀ AI ਮੈਮੋਰੀ ਦਿੰਦਾ ਹੈ, ਅਤੇ ਆਕਸੀਲੈਬਸ ਤੁਹਾਡੇ ਏਜੰਟਾਂ ਨੂੰ ਖੋਜ ਲਈ ਲਾਈਵ ਵੈੱਬ ਡੇਟਾ ਦਿੰਦਾ ਹੈ। ਲਿੰਕ 🐘 ਹਾਥੀ-ਗੋਲਡਫਿਸ਼: google/elephant-goldfish-model 🧠 ਹੋਂਚੋ: go.honcho.dev (ਕੋਡ ਜੈਕ $10 ਕ੍ਰੈਡਿਟ ਲਈ) 🍎 BAMA: daveebbelaar/bama 📊 ਓਪਨ UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io ($5 ਮੁਫ਼ਤ ਕ੍ਰੈਡਿਟ + JACK ਕੋਡ ਦੇ ਨਾਲ 20% ਦੀ ਛੋਟ) 📅 ਪੋਸਟਿਜ਼: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 ਹਾਲਮਾਰਕ: Nutlope/hallmark 🎬 ਮੋਸ਼ਨਸਾਈਟਸ: Motionsites-AI/claude-skill 📈 ਏਜੰਟਿਕ OS + ਪੂਰਾ ਕੋਰਸ ਅਤੇ ਕਮਿਊਨਿਟੀ: bit.ly 🗣️ ਗਲੇਡੋ: get.glaido.com 🤖 ਕਲੌਡ: claude.ai ਸਟੈਂਪਸ 0:00 - ਜਾਣ-ਪਛਾਣ 0:28 - ਪੱਧਰ 1: ਹਾਥੀ-ਗੋਲਡਫਿਸ਼ 2:52 - ਬੋਨਸ: ਹੋਂਚੋ 4:58 - ਪੱਧਰ 2: ਓਪਨ UI 6:20 - ਪੱਧਰ 3: BAMA 9:18 - ਪੱਧਰ 4: ਕਲੌਡ-ਸੀਓ 11:33 - ਬੋਨਸ: ਆਕਸੀਲੈਬਸ 12:58 - ਪੱਧਰ 5: ਹਾਲਮਾਰਕ 14:48 - ਪੱਧਰ 6: ਮੋਸ਼ਨ ਸਾਈਟਸ 15:58 - ਪੱਧਰ 7: ਪੋਸਟਿਜ਼ 17:25 - ਪੱਧਰ 8: ਏਆਈ-ਜੌਬ-ਸਰਚ 18:47 - ਪੱਧਰ 9: ਡੀਪ ਟਿਊਟਰ 19:33 - ਆਉਟਰੋ ▶️ ਅੱਗੇ ਦੇਖੋ 7 ਮੁਫ਼ਤ ਰਿਪੋਜ਼ ਜੋ ਕਲੌਡ ਨੂੰ ਇੰਨਾ ਵਧੀਆ ਬਣਾਉਂਦੇ ਹਨ ਕਿ ਇਹ ਗੈਰ-ਕਾਨੂੰਨੀ ਲੱਗਦਾ ਹੈ: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [pl] 9 repozytoriów na GitHubie, które są teraz hitem (przetestowałem każde z nich) 📈 Agentic OS + Społeczność: bit.ly 🩵 Darmowa umiejętność: bit.ly 👋 Cześć Hej, jestem Jack. Zbudowałem i sprzedałem swój ostatni startup technologiczny z miliardem klientów, a teraz buduję startupy AI i dzielę się tym, co działa. Rzuciłem pracę w korporacji, żeby założyć startup i od tamtej pory nie żałuję. 💬 Przegląd Dziewięć repozytoriów GitHub, które pomogą Ci tworzyć lepsze aplikacje, tworzyć piękne strony internetowe, znajdować pracę i szybciej się uczyć. Przetestowałem każde z nich, a link do każdego znajduje się poniżej. Elephant-Goldfish pomaga Claude'owi wyłapać błędy, zanim napisze kod. BAMA tworzy natywne aplikacje na Maca. Otwarty interfejs użytkownika przekształca odpowiedzi AI w interaktywne wykresy i pulpity nawigacyjne. Claude-Seo audytuje Twoją stronę internetową i przedstawia plan jej ulepszenia. Postiz umieszcza Twoje posty w mediach społecznościowych w jednym kalendarzu. ai-job-search dopasowuje oferty pracy do Twojego CV i pomaga dostosować aplikacje. DeepTutor zamienia plik PDF w korepetytora, który wyjaśnia trudne kwestie i przeprowadza testy. Następnie porównuję Hallmark i MotionSites z tym samym briefem, aby sprawdzić, który daje najlepszy wynik. Dodatkowo, Honcho zapewnia pamięć AI na przestrzeni sesji, a Oxylabs udostępnia agentom dane online na żywo do celów badawczych. Linki 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (kod jack na 10 USD) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (5 USD darmowego kredytu + 20% zniżki z kodem JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Pełny kurs i społeczność: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Znaczki 0:00 - Wprowadzenie 0:28 - Poziom 1: Słoń-Złota Rybka 2:52 - Bonus: Honcho 4:58 - Poziom 2: Otwarty interfejs użytkownika 6:20 - Poziom 3: BAMA 9:18 - Poziom 4: claude-seo 11:33 - Bonus: Oxylabs 12:58 - Poziom 5: Hallmark 14:48 - Poziom 6: MotionSites 15:58 - Poziom 7: Postiz 17:25 - Poziom 8: ai-job-search 18:47 - Poziom 9: DeepTutor 19:33 - Outro ▶️ Oglądaj dalej 7 darmowych repozytoriów, dzięki którym Claude jest tak dobry, że aż wydaje się nielegalny: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [pt-BR] 9 Repositórios do GitHub que estão bombando agora (Eu testei todos) 📈 Agentic OS + Comunidade: bit.ly 🩵 Skill gratuita: bit.ly 👋 Olá! Olá, sou o Jack. Criei e vendi minha última startup de tecnologia com milhões de clientes. Agora, estou criando startups de IA e compartilhando coisas que funcionam. Larguei meu emprego corporativo para me dedicar à minha startup e nunca me arrependi. 💬 Visão geral Nove repositórios do GitHub que ajudam você a criar aplicativos melhores, sites incríveis, encontrar empregos e aprender mais rápido. Testei todos eles e os links estão abaixo. O Elephant-Goldfish ajuda o Claude a detectar erros antes de escrever o código. O BAMA cria aplicativos nativos para Mac. O Open UI transforma respostas de IA em gráficos e painéis interativos. O claude-seo audita seu site e fornece um plano para melhorá-lo. O Postiz reúne suas postagens de redes sociais em um único calendário. O ai-job-search encontra vagas que combinam com seu currículo e ajuda a personalizar suas candidaturas. O DeepTutor transforma um PDF em um tutor que explica as partes mais difíceis e faz perguntas. Em seguida, comparei o Hallmark e o MotionSites com o mesmo briefing de website para ver qual deles oferece o melhor resultado. Além disso, o Honcho fornece memória para sua IA entre sessões, e o Oxylabs oferece aos seus agentes dados da web em tempo real para pesquisa. Links 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (use o código jack para ganhar US$ 10 de crédito) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (US$ 5 de crédito grátis + 20% de desconto com o código JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + Curso Completo e Comunidade: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Marcadores 0:00 - Introdução 0:28 - Nível 1: Elephant-Goldfish 2:52 - Bônus: Honcho 4:58 - Nível 2: Open UI 6:20 - Nível 3: BAMA 9:18 - Nível 4: claude-seo 11:33 - Bônus: Oxylabs 12:58 - Nível 5: Hallmark 14:48 - Nível 6: MotionSites 15:58 - Nível 7: Postiz 17:25 - Nível 8: ai-job-search 18:47 - Nível 9: DeepTutor 19:33 - Encerramento ▶️ Assista a seguir 7 repositórios gratuitos que tornam o Claude tão bom que parece ilegal: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [ru] 9 GitHub-репозиториев, которые сейчас на хайпе (я протестировал каждый) 📈 Agentic OS + Сообщество: bit.ly 🩵 Бесплатный навык: bit.ly 👋 Привет! Привет, меня зовут Джек. Я создал и продал свой последний технологический стартап с миллиардом клиентов, а теперь создаю стартапы в области ИИ и делюсь тем, что работает. Я уволился с корпоративной работы, чтобы заняться своим стартапом, и с тех пор ни разу не пожалел об этом. 💬 Обзор Девять репозиториев GitHub, которые помогут вам создавать лучшие приложения, красивые веб-сайты, находить работу и учиться быстрее. Я протестировал каждый из них, и все ссылки ниже. Elephant-Goldfish помогает Клоду выявлять ошибки до написания кода. BAMA создает нативные приложения для Mac. Open UI превращает ответы ИИ в интерактивные диаграммы и панели мониторинга. claude-seo проводит аудит вашего веб-сайта и предлагает план его улучшения. Postiz объединяет ваши публикации в социальных сетях в одном календаре. ai-job-search подбирает вакансии в соответствии с вашим резюме и помогает адаптировать ваши заявки. DeepTutor превращает PDF-файл в обучающий инструмент, который объясняет сложные моменты и проверяет ваши знания. Затем я сравнил Hallmark и MotionSites с одним и тем же веб-сайтом, чтобы определить, какой из них даст наилучший результат. Кроме того, Honcho предоставляет вашему ИИ память между сессиями, а Oxylabs — данные веб-сайта в режиме реального времени для исследований. Ссылки 🐘 Elephant-Goldfish: google/elephant-goldfish-model 🧠 Honcho: go.honcho.dev (код jack для получения 10 долларов кредита) 🍎 BAMA: daveebbelaar/bama 📊 Open UI: thesysdev/openui 🔎 claude-seo: AgriciDaniel/claude-seo 🌐 Oxylabs: oxylabs.io (5 долларов бесплатного кредита + скидка 20% с кодом JACK) 📅 Postiz: gitroomhq/postiz-app 💼 ai-job-search: ai-job-search-dev/ai-job-search 🎓 DeepTutor: HKUDS/DeepTutor 🎨 Hallmark: Nutlope/hallmark 🎬 MotionSites: Motionsites-AI/claude-skill 📈 Agentic OS + полный курс и сообщество: bit.ly 🗣️ Glaido: get.glaido.com 🤖 Claude: claude.ai Stamps 0:00 - Вступление 0:28 - Уровень 1: Слон-Золотая Рыбка 2:52 - Бонус: Honcho 4:58 - Уровень 2: Open UI 6:20 - Уровень 3: BAMA 9:18 - Уровень 4: claude-seo 11:33 - Бонус: Oxylabs 12:58 - Уровень 5: Hallmark 14:48 - Уровень 6: MotionSites 15:58 - Уровень 7: Postiz 17:25 - Уровень 8: ai-job-search 18:47 - Уровень 9: DeepTutor 19:33 - Заключение ▶️ Смотреть далее 7 бесплатных репозиториев, которые делают Claude настолько хорошим, что кажется незаконным: youtu.be #GitHub #ClaudeCode #OpenSource #AITools #Claude [ta] தற்போது வைரலாகும் 9 GitHub களஞ்சியங்கள் (ஒவ்வொன்றையும் நானே சோதித்தேன்) 📈 ஏஜென்டிக் ஓஎஸ் + சமூகம்: bit.ly 🩵 இலவசத் திறன்: bit.ly 👋 வணக்கம் நான் ஜாக். எனது கடைசி தொழில்நுட்ப ஸ்டார்ட்அப்பை உருவாக்கி, கோடிக்கணக்கான வாடிக்கையாளர்களுடன் அதை விற்றேன். இப்போது நான் AI ஸ்டார்ட்அப்களை உருவாக்கி வருகிறேன், மேலும் சிறப்பாகச் செயல்படும் விஷயங்களைப் பகிர்ந்து கொள்கிறேன். எனது ஸ்டார்ட்அப்பைத் தொடங்குவதற்காக எனது கார்ப்பரேட் வேலையை விட்டுவிட்டேன், அன்று முதல் திரும்பிப் பார்த்ததே இல்லை. 💬 கண்ணோட்டம் சிறந்த செயலிகளை உருவாக்கவும், அழகான வலைத்தளங்களை உருவாக்கவும், வேலைகளைக் கண்டறியவும், மேலும் வேகமாக கற்றுக்கொள்ளவும் உதவும் ஒன்பது கிட்ஹப் ரெப்போக்கள். நான் ஒவ்வொன்றையும் சோதித்தேன், ஒவ்வொரு இணைப்பும் கீழே கொடுக்கப்பட்டுள்ளது. எலிஃபண்ட்-கோல்டுஃபிஷ், கிளாட் கோட் எழுதுவதற்கு முன்பே தவறுகளைக் கண்டறிய உதவுகிறது. பாமா நேட்டிவ் மேக் செயலிகளை உருவாக்குகிறது. ஓப்பன் யுஐ, AI பதில்களை ஊடாடும் விளக்கப்படங்களாகவும் டாஷ்போர்டுகளாகவும் மாற்றுகிறது. கிளாட்-எஸ்இஓ உங்கள் வலைத்தளத்தை தணிக்கை செய்து, அதை மேம்படுத்துவதற்கான திட்டத்தை வழங்குகிறது. போஸ்டிஸட் உங்கள் சமூக ஊடகப் பதிவுகளை ஒரே காலண்டரில் வைக்கிறது. ai-job-search உங்கள் சுயவிவரக் குறிப்புக்கு (CV) ஏற்ற வேலைகளைத் தேர்ந்தெடுத்து, உங்கள் விண்ணப்பங்களைத் தனிப்பயனாக்க உதவுகிறது. DeepTutor ஒரு PDF கோப்பை, கடினமான பகுதிகளை விளக்கி வினாக்கள் கேட்கும் ஒரு பயிற்றுவிப்பாளராக மாற்றுகிறது. பின்னர், எது சிறந்த முடிவைத் தருகிறது என்பதைப் பார்க்க, அதே இணையதளச் சுருக்கத்திற்கு எதிராக Hallmark மற்றும் MotionSites-ஐ நான் ஒப்பிட்டுப் பார்த்தேன். மேலும், Honcho உங்கள் AI-க்கு அமர்வுகள் முழுவதும் நினைவாற்றலை வழங்குகிறது, மற்றும் Oxylabs உங்கள் முகவர்களுக்கு ஆராய்ச்சிக்காக நேரடி இணையத் தரவை வழங்குகிறது. இணைப்புகள் 🐘 எலிஃப

"l' #IA va faire qu'on ne réfléchit plus, et donc rendre con !" - Ah. Si tu l'dis !? x.com Un prof personnel à la maison ! HKUDS/DeepTutor #DeepTutor #AI

VideoRAG: Chat with Your Videos (3.4k stars) What it does All text below is quoted directly from the project README. >[KDD'2026] "VideoRAG: Chat with Your Videos" Key Features: - Drag & Drop Upload: Simply drag video files into Vimo - Smart Conversations: Ask questions in natural language - Multi-Format Support: Works with MP4, MKV, AVI, and more - Cross-Platform: Available on macOS, Windows, and Linux - Extreme Long Videos: Process videos up to hundreds of hours - Multi-Video Analysis: Compare and analyze multiple videos simultaneously - Advanced Retrieval: Find specific moments and scenes with precision - Export Capabilities: Save insights and references for later use Repository HKUDS/VideoRAG Repository stats |Metric|Value| |:-|:-| |Stars|3,436| |Forks|477| |Last release|None| |Languages|Python 57.6%, TypeScript 40.9%, CSS 0.7%| |Contributors|6| |AI contributor accounts (Claude / Codex / Cursor / Copilot)|None| |AI co-authored commits|None| Main contributor |Field|Value| |:-|:-| |Name|Xubin Ren| |GitHub|@Re-bin| |Commits to this repo|55 (top contributor)| |Bio|Ph.D. student of the University of Hong Kong @HKUDS| |Company / Location|The University of Hong Kong| |Followers|629| ────────────────────────────── 📌 How many great GitHub repos have you discovered and completely forgotten about? Don't let another useful tool disappear into your Reddit feed. Get the community's best repos, organized by category, plus key takeaways from real developer discussions, delivered to your inbox twice a week. 📬 Never lose track of great repos again. Subscribe free.