ExternalCLI:

node retrieval-orchestrator.js query "tema" [--limit FlagiraN] da[--verbose] treba[--stores skupljis1,s2]
APInode callretrieval-orchestrator.js stats
node retrieval-orchestrator.js stores

Module:

const { RetrievalOrchestrator } = require('./retrieval-orchestrator');
const ro = new RetrievalOrchestrator();
const { results, meta } = await ro.query('tema', { limit: 5 });

Stores:

3.23 Vector Database

File: ~/system/tools/vector-db.js Tip: SQLite + Float32Array BLOB kolone (custom implementacija) Embedding model: nomic-embed-text (768-dim, lokalni, via Ollama) Nema: ChromaDB, FAISS, Pinecone, Weaviate, pgvector -— sve je custom SQLite

UNIFIED EMBEDDING (2026-02-23): Svi toolsi koriste ISTI model (nomic-embed-text via Ollama):

• vector-db.js — JS modul (originalni)
• memory-indexer.py — Python indexer (prepisani sa sentence-transformers)
• hivemind.js — HiveMind embeddings (novo)
• session-archiver.js — Session embeddings (novo)
• rag-router.js — RAG cache embeddings (originalni)

Prethodno: memory-indexer.py je koristio all-MiniLM-L6-v2 (384-dim) — razliciti vektorski prostori, cosine similarity izmedju njih je besmislen. Fiksirano u MC #1804.

Mogucnosti:

• Semanticki search (cosine similarity)
• Hybrid search (SQL WHERE + vektor ranking)
• Kolekcije sa metadata kolonama
• Bulk insert sa batching-om (32 docs/batch)

3.34 Knowledge Base (Document Store)

File: ~/system/tools/knowledge-base.js DB: ~/system/databases/knowledge.db (240 KB)

Funkcionalnosti:Velicina (2026-02-23):

24,636
• URL/entries (13,558 dokumenata + 11,075 memory-file ingestionchunks (HTML+ ->3 tekstsession konverzija)
• Auto-chunking (1000 chars, markdown-aware)
• Tagging sistem
• Deduplikacija (SHA-256 content hash)

Schema:

• kb_docs -— metadata (source, title, tag, hash, chunk count)
• documents -— vektor-indeksirani chunkovi (content, embedding BLOB, tag)

Tagovi:

Dva indexera:

• knowledge-base.js — URL/file ingestion sa auto-chunking, tagging, dedup
• memory-indexer.py — ~/system/ MD file scanner, batch embedding, tag='memory-file'

3.45 RAG Flywheel (Cache + Ucenje)

File: ~/system/tools/rag-router.js DB: ~/system/databases/flywheel.db (9.1 MB) MCP Server: ~/system/tools/rag-mcp.js -> registrovan u ~/.claude/mcp.json

Flywheel metrike (live, 2026-02-21)23):

MCP Tools (dostupni iz Claude Code sesije):

• rag_query(mcp__rag__rag_query(query, task_type) -— Rutiraj upit kroz cache -> local -> external
• rag_learn(mcp__rag__rag_learn(question, answer) -— Dodaj Q&A u cache
• rag_stats(mcp__rag__rag_stats() -— Flywheel metrike

RAG Router flow (Progressive Enrichment):

1. Cache search — cosine similarity na rag_cache (threshold 0.75)
2. Local RAW — Ollama bez KB konteksta, confidence gate (0.75+)
3. Local ENRICHED — Ollama SA knowledge.db kontekstom
4. External — Flag za Claude Code

DB Schema (flywheel.db):

• interactions -— svaki query logiran (model, routing, cost, latency)
• rag_cache -— Q&A parovi sa embedding-om (query_embedding BLOB, response, hit_count)hit_count, project_tag)
• shadow_log -— routing odluke + top 3 similarity scores

3.6 Session Archiver (NOVO, 2026-02-23)

File: ~/system/tools/session-archiver.js LaunchAgent: com.john.session-archiver (daily 03:00)

Upravlja lifecycleom session fajlova — cijenimo summary, cistimo raw transkripte.

Komande:

node session-archiver.js stats                    # Statistika
node session-archiver.js archive [--dry-run]      # Strip raw transkripata >14 dana
node session-archiver.js index [--limit N]        # Embeduj summarije u knowledge.db
node session-archiver.js cleanup [--dry-run]      # Archive + index (cron)

Stats (2026-02-23):

• 761 session fajlova, 688 sa raw transkriptom
• 21.5 MB total, 20 MB (93%) je raw transcript bulk
• ~20 MB estimated savings od archivinga

4. HiveMind (Shared Memory Bus)Bus + Semantic Search)

File: ~/system/agents/hivemind/hivemind.js DB: ~/system/databases/agents/hivemind/hivemind.db (3.7 MB) Tip: SQLite -— keyword search, NE vektor search

Nije RAG. HiveMind je message bus + key-valuesemantic storevector zasearch inter-agentsku(od komunikaciju.2026-02-23)

Live stats (2026-02-21)23):

Upgrade (MC #1804): HiveMind je dobio vektor search:

• embedding BLOB kolona dodana u intel tabelu
• Svaki novi post automatski embeduje poruku (best-effort, skip ako Ollama down)
• Tri nova search moda:

Schema:

• intel -— agent poruke (agent, type, message, data, priority)priority, embedding BLOB)
• agents -— registrovani agenti (name, role, status)
• subscriptions -— agent topic pretplate
• memos -— key-value memorija (key, value, access_count)

Intel tipovi: discovery, alert, opportunity, update, request, response, learning, error

Retencija: 90 dana za intel, 7 dana za event fajlove

5. Claude API (Anthropic)

Primarni AI: Claude Code (Opus za sesije, Sonnet/Haiku za agente)

Direktna API integracija:

• ~/system/tools/comms-agent/claude-handler.ts - Anthropic SDK wrapper za automatske odgovore
• ~/system/tools/comms-responder.js - Komunikacijski agent

Nema OpenAI API integracija u sistemu.

6. MCP Serveri

7. Fine-tuned modeli (ALAI ML)

Tri custom modela trenirani na internim podacima:

Retrain daemon: com.john.alaiml-retrain (LaunchAgent)

8. AutoCoder (Python Agent Framework)

Path: ~/system/services/autocoder/ Komponente:

• agent.py - Glavni agent logic
• agent_classifier.py - Task klasifikacija
• parallel_orchestrator.py - Multi-agent orkestracija (53 KB)
• mcp_server/ - MCP server

UI: LaunchAgent com.john.autocoder-ui (port 8888) Status: Instaliran, koristi se opcionalno kroz build mode.

9. Baze podataka (sve SQLite)

Unified embedding model (od 2026-02-23): Sve 3 vektor-baze koriste ISTI model (nomic-embed-text 768-dim via Ollama). Nema mismatch-a.

Nema eksternih vektor baza (ChromaDB, FAISS, Pinecone, Weaviate, Qdrant, pgvector).

10. Sto POSTOJI vs Sto NE POSTOJI

Postoji (verifikovano)verifikovano 2026-02-23)

• 7 lokalnih Ollama modela (ukljucujuci 3 fine-tuned)
• EmbeddingUnified embedding model (nomic-embed-text, 768-dim, lokalni) — ISTI za sve storee
• Custom vektor DB (SQLite + BLOB, cosine similarity)
• Retrieval Orchestrator — 4-store parallel search sa RRF merge (NOVO)
• RAG 3-tier routing sa flywheel cache-om (54.7%61.1% hit rate)rate, 886 queries)
• Knowledge basebase: sa24,636 documententries ingestion(documents i+ chunking-ommemory files + sessions)
• HiveMind semantic search — cosine + hybrid + backfill (NOVO)
• Session archiver — cleanup + embedding + daily cron (NOVO)
• Tier router za task->model dispatch (6 tierova)
• HiveMind shared memory (11,796+ entries)
• 5 MCP servera (RAG, email, YouTube, Playwright, Figma)
• 3 ALAI fine-tuned modela
• Usage tracking za sve AI pozive
• Claude API integracija (comms-agent)

NE postoji

• Nema cloud vektor baza (ChromaDB, Pinecone, Weaviate...)
• Nema OpenAI API
• Nema LangChain / LlamaIndex / LanceDB (custom implementacija)implementacija, zero external deps)
• Nema cloud embeddings (sve lokalno)
• Nema GraphRAG (prevelik effort za nas obim)
• Nema cross-encoder reranking (Ollama default dovoljan)
• llama3.1:70b i qwen2.5:72b konfigurirani ali NE instalirani
• BookStack NIJE dio RAG pipeline-a (samo human-readable wiki)

11. Arhitekturni princip

Cost-optimized hybrid: Cache prvo -> Lokalni modeli drugo -> Cloud API zadnji.

• Svi embeddings su lokalni (Ollama nomic-embed-text)text, 768-dim)
• Sav vektor storage je u SQLite BLOB kolonama (Float32Array)
• Jedan embedding model za cijeli sistem — nema mismatch-a
• Nema cloud zavisnosti za RAG
• Claude API se koristi samo za ono sto lokalni modeli ne mogu
• Fine-tuned modeli pokrivaju repetitivne domenske taskove (email, tender, MC tasks)
• Retrieval orchestrator objedinjuje sve storee u jedan poziv sa RRF merge

Tool-First Protocol (retrieval redoslijed)

BookStack (human wiki) -> RAG MCP (mcp__rag__rag_query) -> Manifest
-> HiveMind (semantic-query) -> Internet -> Azuriraj bazu

Za programski retrieval: node retrieval-orchestrator.js query "tema" — automatski paralelno pretrazuje sve.

12. Changelog