Pembuat Peta Pengetahuan
Ubah pengetahuan yang berantakan jadi jaringan utuh
Knowledge fragment pit
Ever feel —— read banyak book, attend banyak class, feel learn banyak, tapi actual need use, head all fragments, can't connect.
E.g. learn machine learning, know decision tree, know neural network, tapi relation apa? Use scenario mana jadi yang? Why Random Forest better single decision tree? These "knowledge between connection" real valuable stuff, most learn method ga help build.
Toss learning content ke OpenClaw, auto extract core concept dan relation, generate visual knowledge graph. Not pretty-but-useless graph, real "see" knowledge overview structured network.
Support output Mermaid, Markdown mind-map format, direct import Obsidian atau other tool. Final exam, new field learn systematic, super useful —— one graph worth three-day note flip.
3 knowledge graph prompt, copy langsung pake
From course organize to note library analyze, choose sesuai butuh.
Organize course content jadi knowledge graph:
[Paste course note / textbook table-of-content / slide content]
Requirement:
1. Extract semua core concept (max 30)
2. Annotate concept relation type (include, depend, compare, evolve)
3. Output Mermaid graph TD format
4. Setiap core concept attach satu-sentence explain
5. Mark 3 easily-mix concept pair, explain differenceAnalyze book content, generate mind-map structure:
Book: [book name]
Reading note: [paste note atau chapter summary]
Requirement:
1. Extract book 5-8 core viewpoint
2. Each viewpoint list 2-3 support evidence
3. Mark relation logic viewpoint between (cause-effect / parallel / progressive)
4. Output indent Markdown list format
5. Final summary: book worth remember 3 takeawayGive Obsidian note library content, bantu analyze:
[Paste note filename list, atau export note content]
Analysis requirement:
1. Identify semua main knowledge domain (cluster)
2. Find "knowledge island" ——content isolated no relation other note
3. Discover hidden cross-domain connection: knowledge look unrelated actually connected?
4. Suggest need add "bridge note" ——concept not write tapi can connect current knowledge
5. Output global knowledge graph (Mermaid format)
Arrange domain high-to-low knowledge density.Recommend config
Task type: knowledge graph build / concept relation analyze
Recommend model: Claude Opus 4.6 (deep understand concept relation)
Backup model: DeepSeek V3 (quick simple mind-map process)
Context advice: give all content once, avoid batch split cause break
Output format: Mermaid graph TD (can direct render)
Temperature set: 0.3 (reduce concept relation "creative make-up")Knowledge graph: OpenClaw vs manual organize
DIY also possible, efficiency difference too big.
- One course content 5 min generate complete knowledge graph
- Auto discover concept relation you miss
- Support Mermaid / Markdown format, direct import Obsidian
- Can iterate: add new content, graph auto expand
- One course draw decent knowledge graph need half-day minimum
- Easy miss concept relation hidden
- XMind tool manual drag, format adjust waste half time
- Content update, whole graph redraw