VideoLearn imports videos from Bilibili or your local files, transcribes speech in real time, and answers questions grounded in the current playback position. Each conversation is distilled into a structured note.
Try it nowPaste a Bilibili link or upload a local video; downloading and processing runs automatically
Speech is converted to text in real time; ask questions grounded in the video context
AI distills the key points and concepts from the conversation into a structured note
A complete learning loop from import to notes
Paste a Bilibili link and authorize via QR scan on your phone. BV IDs and b23.tv short links are both supported. yt-dlp runs server-side to bypass IP-based anti-scraping.
Cloud-based recognition backed by Zhipu GLM-ASR, sliced into precise 30-second segments. On-demand transcription fills in missing segments while you chat — no waiting for the whole video to finish.
The assistant builds a precise "from last question to current playback" context window, using transcript text plus key frames to answer — it genuinely understands what you're watching.
Mark key moments with one click during playback. Ask questions tied to a specific bookmark to quickly locate and revisit important parts.
From your conversation with the assistant, AI distills headings, key points, and core concepts into a structured JSON note. No manual cleanup — you finish learning with a complete note.
Answers stream back in real time over Server-Sent Events, word by word. Markdown rendering and code highlighting keep the interaction smooth and natural.
Every video session ends with something to keep
Import a Bilibili tutorial or open lecture; ask "why is this code written this way" or "how is this formula derived" while you watch, and get chapter-level notes automatically.
Upload a recorded lecture or meeting; the assistant transcribes it in full. Mark the important moments with bookmarks, ask questions to go deeper, and finish with a complete note.
For team-internal tech talks or conference talks, the assistant extracts the core approach, architecture, and technical details — no need to re-watch repeatedly.
Teachers upload class recordings; the assistant helps check coverage. Students use Q&A to fill gaps in what they didn't catch in class — personalized learning on demand.
A modern full-stack architecture — async and performant
Import a video, talk to AI, generate notes automatically. Try it free with demo / demo123.
Visit VideoLearnCrawls trending HuggingFace papers; an LLM produces summaries and categories, refreshed daily with the latest AI progress.
Upload a PDF; the model extracts key information and generates a mind map for deep paper reads and fast technical-document skims.
17 AI agents work together across HR, finance, IT operations, and more — natural-language-driven office workflows.
A transparent proxy for LLM API calls that compresses redundant context. One-line integration cuts spend by 30–60%.
A tiered knowledge database built for AI agents — L0/L1/L2 summaries, MCP integration, and session memory.