LOCAL AI KNOWLEDGE BASE SYSTEM

LOCAL AI KNOWLEDGE BASE SYSTEM NO INTERNET REQUIRED LOCAL AI KNOWLEDGE BASE NO INTERNET REQUIRED LOCAL AI KNOWLEDGE BASE SYSTEM ALL PRACTICAL CONTENT NO FILLER ALL PRACTICAL CONTENT NO USELESS WORDS EVEN COMPUTER NOVICES CAN UNDERSTAND AND FOLLOW HOPE EVERYONE CAN LIKE AND FOLLOW WHILE SAVING THANK YOU OLD IRON! THANK YOU EVERYONE PART 1: THEORY EXPLANATION BEFORE FORMALLY INTRODUCING THE SPECIFIC OPERATION I WANT TO SIMPLY CHAT WITH YOU BRIEFLY ABOUT A FEW SMALL QUESTIONS IN BUILDING A LOCAL KNOWLEDGE BASE INCLUDING WHY BUILD AN AI KNOWLEDGE BASE? WHAT IS THE IMPLEMENTATION PRINCIPLE OF AN AI KNOWLEDGE BASE? WHAT SOFTWARE IS NEEDED TO BUILD A KNOWLEDGE BASE? I THINK UNDERSTANDING THESE QUESTIONS IS MORE IMPORTANT THAN LEARNING HOW TO OPERATE AND BUILD A KNOWLEDGE BASE WHY BUILD AN AI KNOWLEDGE BASE? LET'S LOOK AT THE FIRST QUESTION WHY DO WE NEED TO BUILD AN AI KNOWLEDGE BASE? WITHOUT AI, OUR INFORMATION IS OFTEN STORED IN A FOLDER OR STORED IN VARIOUS NOTE-TAKING SOFTWARE SEARCHING FOR INFORMATION RELIES PURELY ON MEMORY OR THE SOFTWARE'S SEARCH FUNCTION NOT ONLY IS THE EFFICIENCY LOW, BUT IT IS ALSO EASY TO FORGET THE MOST IMPORTANT THING IS THAT WHAT YOU FIND IS MERELY THE ORIGINAL DOCUMENT NOT THE DIRECT ANSWER TO YOUR QUESTION AND AFTER HAVING AN AI KNOWLEDGE BASE AI WILL FIRST ANALYZE AND UNDERSTAND ALL YOUR DOCUMENT INFORMATION WHEN YOU SEARCH IT CAN QUICKLY UNDERSTAND YOUR INTENTION AND MATCH THE MOST ACCURATE CONTENT FROM THE KNOWLEDGE BASE MATCH THE MOST ACCURATE CONTENT WHAT IS THE PRINCIPLE OF AN AI KNOWLEDGE BASE? IT'S ACTUALLY VERY SIMPLE YOU CAN UNDERSTAND IT AS A "SEMANTIC SEARCH SYSTEM THAT CAN UNDERSTAND MEANING" THE ENTIRE PROCESS IS APPROXIMATELY DIVIDED INTO THREE STEPS STEP 1: CONVERT DOCUMENTS INTO VECTORS WHEN YOU IMPORT A PDF, MARKDOWN NOTE, OR WORD DOCUMENT THE SYSTEM WILL FIRST BREAK DOWN THE TEXT CONTENT INTO MANY SMALL SEGMENTS TEXT SEGMENTS NUMERIC VECTORS THEN USE SOMETHING CALLED "EMBEDDING MODEL" (EMBEDDING MODEL) CONVERT THESE TEXTS INTO A STRING OF NUMBERS, WHICH IS "VECTOR" CONVERTED INTO A FORM THAT COMPUTERS CAN UNDERSTAND THE FUNCTION OF THIS STEP IS TO ALLOW COMPUTERS TO "UNDERSTAND" THE SEMANTIC INFORMATION OF EACH PIECE OF TEXT STEP 2: SEMANTIC SEARCH WHEN YOU ASK A QUESTION, FOR EXAMPLE, "DOES THIS DOCUMENT MENTION CRYPTOGRAPHIC ALGORITHMS?" THE SYSTEM WILL NOT DO A SIMPLE KEYWORD MATCH INSTEAD, IT WILL ALSO CONVERT YOUR QUESTION INTO A VECTOR AND THEN FIND THE MOST SIMILAR SEGMENT OF CONTENT AMONG ALL DOCUMENT VECTORS THIS IS LIKE UNDERSTANDING IN THOUSANDS OF DOCUMENTS INSTEAD OF KEYWORD MATCHING AI USES "UNDERSTANDING" INSTEAD OF "KEYWORDS" TO HELP YOU FIND ANSWERS STEP 3: LARGE MODEL GENERATES ANSWERS AFTER FINDING THE RELEVANT CONTENT THE SYSTEM WILL SEND THESE CONTENT SEGMENTS TO THE LARGE LANGUAGE MODEL (LLM) SUCH AS LLAMA, Qwen, Deepseek, Gemma, AND OTHER OPEN-SOURCE MODELS THE MODEL WILL GENERATE A NATURAL LANGUAGE RESPONSE BASED ON THIS CONTENT IN SUMMARY, THE ENTIRE PROCESS IS: TEXT -> VECTORIZATION -> SEMANTIC SEARCH -> MODEL RESPONSE THEREFORE, THE CORE OF A LOCAL AI KNOWLEDGE BASE IS THAT IT NOT ONLY HELPS YOU "FIND INFORMATION" BUT ALSO "USE INFORMATION" AND MOST IMPORTANTLY THE ENTIRE PROCESS CAN BE COMPLETED OFFLINE ON YOUR COMPUTER WITHOUT CONNECTING TO ANY EXTERNAL SERVERS DATA IS ALWAYS IN YOUR HANDS WHAT SOFTWARE IS NEEDED TO BUILD AN AI KNOWLEDGE BASE? LET'S FINALLY LOOK AT THE THIRD QUESTION WHAT TOOLS ARE NEEDED TO BUILD AN AI KNOWLEDGE BASE? THE CONSTRUCTION PROCESS IS NOT AS COMPLICATED AS IMAGINED THE ENTIRE SYSTEM IS MAINLY COMPOSED OF TWO PARTS STEP 1: BUILD THE LOCAL MODEL RUNNING ENVIRONMENT THE FIRST PART IS TO BUILD THE RUNNING ENVIRONMENT FOR THE LOCAL MODEL THE FUNCTION OF THIS PART IS EQUIVALENT TO THE "BRAIN" OF THE ENTIRE KNOWLEDGE BASE SPECIFICALLY, IT IS TO DEPLOY OPEN-SOURCE LARGE LANGUAGE MODELS SUCH AS DEEPSEEK, Qwen, ETC. ON THE LOCAL COMPUTER ONLY DEPLOYED FOR LOCAL USE CAN TRULY ACHIEVE OFFLINE AND INTERNET-FREE OPERATION THERE ARE MANY TOOLS AVAILABLE NOW TO HELP US DEPLOY LOCAL MODELS SUCH AS Ollama, LM Studio THIS VIDEO WILL USE Ollama FOR DEMONSTRATION BECAUSE IT DOES NOT REQUIRE SCIENTIFIC INTERNET ACCESS AND CAN EASILY DEPLOY MODELS AND THE NEW VERSION OF Ollama HAS ADDED A CHAT UI THAT CAN DIRECTLY CHAT WITH THE MODEL WHICH IS MORE SUITABLE FOR NOVICE USERS IF YOU ARE NOT A NOVICE AND CAN ACCESS THE INTERNET SCIENTIFICALLY THEN I WOULD MORE RECOMMEND YOU TO USE LM STUDIO ITS FUNCTIONS ARE MORE RICH STEP 2: BUILD THE KNOWLEDGE BASE MANAGEMENT TOOL THE SECOND PART IS TO BUILD THE KNOWLEDGE BASE MANAGEMENT TOOL THIS IS THE KEY COMPONENT THAT ALLOWS AI TO "UNDERSTAND YOUR DOCUMENTS" YOU CAN UNDERSTAND IT AS THE "BRIDGE" CONNECTING "DOCUMENTS" AND "MODELS" THERE ARE MORE SOFTWARE OPTIONS FOR THIS PART SUCH AS AnythingLLM, Cherry Studio, ETC. THIS VIDEO WILL CHOOSE ANYTHINGLLM FOR DEMONSTRATION BECAUSE I TESTED AND FOUND THAT AnythingLLM'S KNOWLEDGE BASE USAGE EFFECT IS MUCH BETTER THAN Cherry Studio PART 2: PRACTICAL OPERATION OK, THE THEORY PART IS DONE NEXT, WE WILL ENTER THE PRACTICAL OPERATION SECTION TO GUIDE YOU HANDS-ON TO BUILD A PURELY OFFLINE LOCAL AI KNOWLEDGE BASE FIRST, WE GO TO Ollama'S OFFICIAL WEBSITE TO DOWNLOAD THE OllAMA CLIENT AFTER ENTERING, CLICK THIS DOWNLOAD BUTTON IT WILL ENTER THE DOWNLOAD PAGE SELECT YOUR SYSTEM TYPE CLICK DOWNLOAD TO INSTALL AFTER INSTALLATION, OPEN THE SOFTWARE YOU CAN SEE THE OllAMA INTERFACE I INTRODUCED OllAMA TO EVERYONE BEFORE IT CAN ONLY CHAT THROUGH TERMINAL COMMANDS IT CAN ONLY CHAT THROUGH TERMINAL COMMANDS THE NEW VERSION OF Ollama HAS ADDED A CHAT UI THAT CAN DIRECTLY CHAT WITH THE MODEL NOT ONLY THAT, THE DOWNLOAD METHOD OF MODELS HAS ALSO BECOME VERY SIMPLE CLICK THE MODEL LIST IN THE LOWER RIGHT CORNER OF THE INPUT BOX YOU CAN SEE THAT IT HAS LISTED SOME COMMONLY USED OPEN-SOURCE MODELS MODELS WITH A CLOUD ICON MEAN THEY ARE CLOUD MODELS PROVIDED BY Ollama CLOUD MODELS REQUIRE INTERNET + OllAMA ACCOUNT AS LONG AS YOU HAVE INTERNET AND REGISTER AN OllAMA ACCOUNT, YOU CAN CHAT DIRECTLY CLOUD MODELS ARE ALL OPEN-SOURCE MODELS WITH VERY LARGE PARAMETERS FOR EXAMPLE, DEEPSEEK V3 671B MODEL QWEN3-CODER'S 480B MODEL THESE MODELS ARE BASICALLY DIFFICULT TO DEPLOY DIRECTLY ON PERSONAL COMPUTERS HOWEVER, THE FOCUS OF THIS ISSUE IS THE CONSTRUCTION OF OFFLINE LOCAL AI KNOWLEDGE BASES SO CLOUD MODELS ARE NOT OUR FOCUS WHAT WE NEED TO LOOK AT ARE THE FOLLOWING MODELS WITH DOWNLOAD ICONS OPENAI'S OPEN-SOURCE GPT-OSS MODELS GOOGLE'S OPEN-SOURCE GEMMA3 MODELS AND DOMESTIC DEEPSEEK AND QWEN MODELS HERE I WILL USE THE 8B DEEPSEEK-R1 MODEL AS AN EXAMPLE FOR DEMONSTRATION STEP 1: DOWNLOAD LOCAL MODELS AFTER DOWNLOADING, YOU CAN RETURN TO THE OllAMA INTERFACE THEN ENTER THE MODEL YOU JUST DOWNLOADED IN THE INPUT BOX TO CHAT IT WILL AUTOMATICALLY DOWNLOAD THIS MODEL WAIT FOR THE MODEL DOWNLOAD TO COMPLETE AFTER THE MODEL DOWNLOAD IS COMPLETE YOU CAN CHAT WITH THIS MODEL LOCALLY LOCAL MODEL CAN BE USED OFFLINE AFTER INSTALLATION EVEN WITHOUT INTERNET, IT CAN BE USED IF YOU WANT TO DOWNLOAD OTHER OPEN-SOURCE MODELS YOU CAN GO TO Ollama'S OFFICIAL WEBSITE CLICK THE MODELS IN THE UPPER LEFT CORNER SEARCH FOR OTHER MODELS HERE BUT THE COMMONLY USED ONES ARE THESE FEW FOR EXAMPLE, IF YOU WANT TO USE DEEPSEEK-R1 14B MODEL COPY THIS COMMAND THEN OPEN THE TERMINAL WINDOW ON YOUR COMPUTER PASTE THE COMMAND YOU CAN PROCEED WITH MODEL DOWNLOAD AFTER DOWNLOADING, YOU CAN RETURN TO THE OllAMA INTERFACE THEN ENTER THE MODEL YOU JUST DOWNLOADED IN THE INPUT BOX TO CHAT STEP 2: INSTALL KNOWLEDGE BASE MANAGEMENT SOFTWARE THE SECOND STEP IS TO DOWNLOAD THE KNOWLEDGE BASE MANAGEMENT SOFTWARE WHICH IS THE ANYTHINGLLM MENTIONED JUST NOW IT IS STILL DIRECTLY DOWNLOADED FROM THE OFFICIAL WEBSITE AFTER INSTALLATION, WE NEED TO PERFORM SOME CONFIGURATION FIRST, IT WILL PROMPT YOU TO SET UP THE LARGE MODEL SERVICE PROVIDER BECAUSE WE USED OllAMA JUST NOW SO WE CHOOSE OllAMA HERE THEN SELECT THE LOCAL MODEL YOU DOWNLOADED FROM THE MODEL LIST I WILL CHOOSE THE 14B DEEPSEEK-R1 MODEL I JUST DOWNLOADED THEN KEEP THE FOLLOWING CONFIGURATION AS DEFAULT AND CLICK NEXT CREATE YOUR FIRST WORKSPACE AND START USING ANYTHINGLLM. FINALLY, IT WILL ASK YOU TO SET THE NAME OF A WORKSPACE YOU CAN ALSO UNDERSTAND THIS WORKSPACE AS ONE OF YOUR KNOWLEDGE BASES AFTER SETTING UP, YOU ENTER THE ANYTHINGLLM WORKSPACE IT PRESENTS A TWO-COLUMN STRUCTURE OVERALL THE LEFT SIDE IS THE WORKSPACE AND CHAT HISTORY LIST THE RIGHT SIDE IS THE CHAT INTERFACE THEN WE CLICK THE SETTINGS BUTTON IN THE LOWER LEFT CORNER CHANGE THE DISPLAY LANGUAGE CONFIGURATION TO CHINESE THEN CLICK LARGE LANGUAGE MODEL (LLM) UNDER AI PROVIDER THIS WAS ACTUALLY CONFIGURED WHEN WE FIRST OPENED THE SOFTWARE IF YOU WANT TO MODIFY IT TO OTHER LARGE MODELS THEN CLICK EMBEDDER UNDER AI PROVIDER CONFIGURE THE EMBEDDING MODEL WE JUST DOWNLOADED EMBEDDER PROVIDER SELECT OllAMA SELECT BGE-M3 IN THE MODEL LIST DON'T FORGET TO CLICK THE SAVE BUTTON IN THE UPPER RIGHT CORNER OTHER SETTINGS CAN BE KEPT AS DEFAULT FOR NOW THEN CLICK NEXT CREATE YOUR FIRST WORKSPACE AND START USING ANYTHINGLLM. FINALLY, IT WILL ASK YOU TO SET THE NAME OF A WORKSPACE YOU CAN ALSO UNDERSTAND THIS WORKSPACE AS ONE OF YOUR KNOWLEDGE BASES AFTER SETTING UP, YOU ENTER THE ANYTHINGLLM WORKSPACE IT PRESENTS A TWO-COLUMN STRUCTURE OVERALL THE LEFT SIDE IS THE WORKSPACE AND CHAT HISTORY LIST THE RIGHT SIDE IS THE CHAT INTERFACE THEN WE CLICK THE SETTINGS BUTTON IN THE LOWER LEFT CORNER THEN CLICK CHAT SETTINGS CHAT MODE HAS TWO OPTIONS HERE ONE IS CHAT, THE OTHER IS QUERY IF YOU CHOOSE CHAT MODE PART OF THE ANSWER YOU GET WILL BE THE LARGE LANGUAGE MODEL'S OWN EXPERIENCE IF YOU CHOOSE QUERY MODE YOUR ANSWER WILL BE COMPLETELY BASED ON THE KNOWLEDGE BASE YOU UPLOADED GENERALLY SET TO QUERY MODE THIS GREATLY IMPROVES THE ACCURACY OF DOCUMENT SEARCH KNOWLEDGE BASE WITH THIS, THE LOCAL AI KNOWLEDGE BASE SYSTEM WE BUILT HAS BEEN SUCCESSFULLY COMPLETED. PURE OFFLINE LOCAL AI KNOWLEDGE BASE PART 1: THEORY EXPLANATION 1. WHY BUILD AN AI KNOWLEDGE BASE? 2. WHAT IS THE IMPLEMENTATION PRINCIPLE OF AN AI KNOWLEDGE BASE? 3. WHAT SOFTWARE IS NEEDED TO BUILD AN AI KNOWLEDGE BASE? PART 2: PRACTICAL OPERATION 1. BUILD LOCAL LARGE MODEL 2. KNOWLEDGE BASE MANAGEMENT SOFTWARE ACTUALLY, IF YOU WANT TO USE THE AI KNOWLEDGE BASE WELL, THERE ARE MANY SKILLS IF THE DATA IN THIS VIDEO IS ACCEPTABLE, I WILL CONTINUE TO PRODUCE VIDEOS ON AI KNOWLEDGE BASE USAGE SKILLS. THIS VIDEO FOCUSES ON ENVIRONMENT CONSTRUCTION AND SOFTWARE INSTALLATION.