Autonomous · AI-native · No build required
VectorMBE
An MBSE substrate that connects your engineering tools into a governed knowledge graph. AI agents draft requirements, synthesize safety assessments, and trace change impact. Engineers review and approve every output.
U.S. Provisional Patent App. No. 64/073,689 — Patent Pending.
How it works
Three interlocking layers
Layer 1
OWL graph as the system of record
Every requirement, component, interface, and verification item is a node in an OWL knowledge graph. Relationships between them are formal and queryable. The graph is the single source of truth — tools feed it, they do not replace it.
Layer 2
Vector embeddings for retrieval
Each node also carries a vector embedding, enabling semantic similarity search across requirements, simulation results, and test evidence. Find analogous past decisions, surface related constraints, and identify reuse opportunities by meaning rather than keyword.
Layer 3
MCP for tool coordination
The Model Context Protocol connects VectorMBE to your existing toolchain. Register the MCP server in Claude Desktop or Cursor and your AI assistant can read from and write to the governed graph. Changes propagate to connected tools automatically.
Capabilities
What VectorMBE does today
Requirements drafting
AI agents extract and draft formal requirements from source documents. Every requirement is linked to its source with full provenance. Engineers review and approve before anything is committed to the graph.
Safety assessment synthesis
Generate structured safety assessments from the graph. Failure modes, hazards, and mitigations are linked to the requirements and components they cover. The graph enforces constraint gates so safety-critical items cannot be bypassed.
Change impact tracing
When a requirement or component changes, the graph traversal identifies everything downstream: interfaces, tests, verification items, and affected safety assessments. Nothing falls through the cracks because the graph knows the dependencies.
Verification evidence tracking
Test results, inspection records, and analysis artifacts are attached to the requirements they verify. The graph surfaces open verification gaps and shows closure status across the program without manual status collection.
Hands-on
Connect via MCP
Register the VectorMBE MCP server in Claude Desktop or Cursor. Run the matching script below, then fully quit and restart the application. Requires vectormbe-mcp on your PATH.
Before you start
vectormbe-mcpmust be on your PATH, or edit the script to use the absolute path to your binary.- After the script runs, fully quit and restart the application.
Cursor — ~/.cursor/mcp.json
python3 << 'EOF'
import json, os
config_path = os.path.expanduser("~/.cursor/mcp.json")
os.makedirs(os.path.dirname(config_path), exist_ok=True)
try:
with open(config_path) as f:
config = json.load(f)
except (FileNotFoundError, json.JSONDecodeError):
config = {}
config.setdefault("mcpServers", {})["vectormbe-runtime"] = {
"type": "stdio",
"command": "vectormbe-mcp",
"args": [],
"env": { "VECTORMBE_LOG_LEVEL": "info" }
}
with open(config_path, "w") as f:
json.dump(config, f, indent=2)
print("Done! Config written to:", config_path)
EOFClaude Desktop — macOS
python3 << 'EOF'
import json, os
config_path = os.path.expanduser("~/Library/Application Support/Claude/claude_desktop_config.json")
os.makedirs(os.path.dirname(config_path), exist_ok=True)
try:
with open(config_path) as f:
config = json.load(f)
except (FileNotFoundError, json.JSONDecodeError):
config = {}
config.setdefault("mcpServers", {})["vectormbe-runtime"] = {
"type": "stdio",
"command": "vectormbe-mcp",
"args": [],
"env": { "VECTORMBE_LOG_LEVEL": "info" }
}
with open(config_path, "w") as f:
json.dump(config, f, indent=2)
print("Done! Config written to:", config_path)
EOFClaude Desktop — Linux / WSL2
python3 << 'EOF'
import json, os
config_path = os.path.expanduser("~/.config/Claude/claude_desktop_config.json")
os.makedirs(os.path.dirname(config_path), exist_ok=True)
try:
with open(config_path) as f:
config = json.load(f)
except (FileNotFoundError, json.JSONDecodeError):
config = {}
config.setdefault("mcpServers", {})["vectormbe-runtime"] = {
"type": "stdio",
"command": "vectormbe-mcp",
"args": [],
"env": { "VECTORMBE_LOG_LEVEL": "info" }
}
with open(config_path, "w") as f:
json.dump(config, f, indent=2)
print("Done! Config written to:", config_path)
EOFClaude Desktop — Windows (PowerShell)
Writes to %APPDATA%\Claude\claude_desktop_config.json. If python is not found, try py -3.
@'
import json, os
from pathlib import Path
appdata = os.environ.get("APPDATA")
if not appdata:
raise SystemExit("APPDATA is not set.")
config_path = Path(appdata) / "Claude" / "claude_desktop_config.json"
config_path.parent.mkdir(parents=True, exist_ok=True)
try:
config = json.loads(config_path.read_text(encoding="utf-8"))
except (FileNotFoundError, json.JSONDecodeError):
config = {}
config.setdefault("mcpServers", {})["vectormbe-runtime"] = {
"type": "stdio",
"command": "vectormbe-mcp",
"args": [],
"env": { "VECTORMBE_LOG_LEVEL": "info" }
}
config_path.write_text(json.dumps(config, indent=2), encoding="utf-8")
print("Done! Config written to:", config_path)
'@ | python -Research foundation
Built on the DoD digital engineering standard
VectorMBE's architecture is grounded in the SERC Handbook on Digital Engineering with Ontologies v2.0, published by the Systems Engineering Research Center under the DoD Office of the Under Secretary of Defense for Research and Engineering. The handbook establishes formal ontologies as the semantic backbone for DoD digital engineering programs. VectorMBE applies those principles in a commercial platform.
Ready to connect your toolchain?
Talk to us about your program. We work with aerospace, automotive, defense, and infrastructure teams.