BISSARA | One Brain. Any AI.

How BISSARA Works

BISSARA Neuro Brain is a structured knowledge system, not an AI model. It provides a searchable, graph-connected knowledge base of Tausug language, grammar, dictionary entries, and cultural knowledge that any AI system can query.

Your AI → BISSARA Connector → Neuro Brain → Knowledge → Context → Your AI Responds

The Neuro Brain contains 1,254 knowledge nodes connected by 4,373 neural pathways across three categories: Grammar (203), Dictionary (471), and AI/Knowledge (580).

Architecture

BISSARA is AI-agnostic. The same knowledge brain works with any model:

┌─────────────────┐
│   AI MODEL      │  Claude, ChatGPT, Gemini, Ollama, Custom
│   (yours)       │
└────────┬────────┘
         │
┌────────▼────────┐
│   CONNECTOR     │  MCP, Python SDK, JS SDK, REST API
└────────┬────────┘
         │
┌────────▼────────┐
│   NEURO BRAIN   │  Retrieval · Search · Graph · Context
└────────┬────────┘
         │
┌────────▼────────┐
│   KNOWLEDGE     │  Grammar · Dictionary · AI/QA
│   1,254 nodes   │  4,373 connections
└─────────────────┘

Data Files

nodes_public.json    # All knowledge nodes with public fields
graph_web.json       # Knowledge graph (nodes + edges)
search_index.json    # Optimized search index
manifest.json        # Brain metadata and stats

Search

Search uses title matching, tag matching, and keyword overlap scoring. Results are ranked by relevance.

// Search algorithm (client-side)
function search(query, index) {
  const q = query.toLowerCase();
  const words = new Set(q.split(/\s+/));

  return index
    .map(entry => {
      const title = entry.title.toLowerCase();
      let score = 0;

      if (q === title) score = 1.0;
      else if (title.includes(q)) score = 0.8;
      else {
        const titleWords = new Set(title.split(/\s+/));
        const overlap = [...words].filter(w => titleWords.has(w));
        score = overlap.length / words.size * 0.6;
      }

      return { ...entry, score };
    })
    .filter(r => r.score > 0)
    .sort((a, b) => b.score - a.score);
}

Knowledge Nodes

Each knowledge node contains:

{
  "id": "uuid-string",
  "title": "Originator focus affixes in tausug",
  "category": "Grammar",       // Grammar | Dictionary | AI
  "language": "Tausug",
  "tags": ["grammar", "rule"],
  "summary": "...",
  "aliases": [],
  "sources": [],
  "related_notes": [],
  "links": [],
  "status": "active"
}

Knowledge Graph

Nodes are connected by edges with different relationship types:

// Edge types
{
  "source": "node-id-1",
  "target": "node-id-2",
  "relation": "title_keyword"    // links_to | related_note |
                                  // shares_tag | title_keyword |
                                  // category_neighbor
}

Use edges to find related concepts, build knowledge trees, or provide context to AI models.

MCP Integration

The BISSARA MCP Server exposes the Neuro Brain as MCP tools for Claude Desktop, VS Code, and compatible clients.

Available Tools

bissara_searchSearch all knowledge

bissara_get_nodeGet a specific node by ID

bissara_neighborsGet connected concepts

bissara_dictionarySearch Tausug dictionary

bissara_grammarSearch Tausug grammar

bissara_contextBuild RAG context for AI

Claude Desktop Configuration

{
  "mcpServers": {
    "bissara-neuro": {
      "command": "python",
      "args": ["/path/to/mcp_server.py"]
    }
  }
}

VS Code Configuration

// .vscode/mcp.json
{
  "servers": {
    "bissara-neuro": {
      "command": "python",
      "args": ["/path/to/mcp_server.py"]
    }
  }
}

Python SDK

from bissara import Bissara

brain = Bissara(data_dir="./data")

# Search
results = brain.search("originator focus")
for r in results:
    print(f"[{r['score']}] {r['title']}")

# Dictionary lookup
words = brain.dictionary("verb")

# Grammar rules
rules = brain.grammar("affix")

# Get specific node
node = brain.get_node("some-node-id")

# Get connected concepts
neighbors = brain.neighbors(node["id"])

# Build AI context
ctx = brain.context("What is the -um- affix?")
print(ctx["context"])  # Ready for LLM prompt

Quick Install

The fastest way to get started is using the interactive installer. It automatically downloads the latest BISSARA Brain data and sets up the SDK of your choice (Python, Node.js, or MCP).

git clone https://github.com/Nasrif30/BISSARA-Neuro-Brain.git
cd BISSARA-Neuro-Brain
./install.sh

Node.js / JavaScript SDK

After running the Quick Install script, you can use the official Node.js SDK in your backend applications.

const { Bissara } = require('bissara-sdk');

// Automatically loads the BISSARA data
const brain = new Bissara();

// Search for any concept
const results = brain.search("verb focus");
console.log(`Found ${results.length} results`);

// Dictionary lookup
const words = brain.dictionary("verb");

// Get connected nodes
const node = brain.getNode("some-id");
const neighbors = brain.getNeighbors(node.id);

// Generate AI context string
const aiContext = brain.context("What is the -um- affix?");
console.log(aiContext.context); // Pass this directly to your LLM!

Local AI with Ollama

Run BISSARA knowledge with a local AI model — no cloud, no API keys, completely free.

# 1. Install Ollama
#    https://ollama.com/download

# 2. Pull a model
ollama pull gemma3

# 3. Download the BISSARA connector
#    (from /downloads page)

# 4. Start BISSARA with Ollama
python bissara_ollama.py

# Result:
# ✓ Ollama detected
# ✓ Neuro Brain loaded (1,254 neurons)
# ✓ Graph ready (4,373 connections)
# ✓ BISSARA connected
#
# Ask: What is the -um- affix in Tausug?

Developer Documentation