A
licenseNot graded
qualityD
maintenanceEnables autonomous orchestration of vector search, knowledge graph queries, and web crawling through a single MCP interface, providing agentic RAG capabilities for AI assistants.
9
MIT
GlassBox OS
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@GlassBox OSShow me the component map and its attributes."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
GlassBox OS
A source-first orchestration and knowledge-indexing toolkit that makes architectural capabilities inspectable through one Python API and MCP surface.
Status: Early-stage research implementation. The repository includes the orchestrator, context/indexing utilities, compatibility MCP adapters, Eco Index experiments, Declarum prototype, tests, and architecture reports. Some component-map entries remain placeholders.
Quick source verification
The non-invasive checks below require only Python 3.11 or newer and do not install packages or modify system configuration:
python -m compileall -q orchestrator mcp-servers mcp_servers scripts tests
python -c "from orchestrator.atomic_ingest import atomic_ingest_session, recover; from orchestrator.discovery_resolver import resolve; print('portable imports: ok')"
python -m orchestrator.cli inspect README.mdFor the full dependency installation and six-part foundation smoke suite, see docs/INSTALLING.md. Review the platform-specific commands before running them: the Termux path includes package-manager and optional ADB steps.
Related MCP server: code-auditor-mcp
Architectural commitment
Declare structure first, validate at boundaries, multi-modal access (human / AI / voice / code) through the same declared layer. Engelbart-style intelligence augmentation in 2026 form.
Invariants:
No neural embedder at any cycle — named-attribute methodology, Claude IS the scorer
All persistence text-format-dumped at commit time (via
gitsqlite+npz2json) — fresh clone = full AI memory, no binary bloatSingle FastMCP-mux process exposes 7 namespaces — survives Claude Code's 60s startup wall + Android's Phantom Process Killer ceiling
Atomic multi-store ingest via POSIX rename — no silent stale fusion
Component map (11 layers)
Layer | Component | Status |
L1 substrate | declarum | placeholder (Stage 1.1) |
L2 cognition | nlke-metacog + nlke-ml + tool_design_assistant | placeholder (Stage 1.1) |
L3 voice modality | voice-graph-rag | Cycle 3 |
L4a inline annotations | claude-context-kit CLAUDE-CTX | Cycle 2 |
L4b folder maps | aria-personal/ctx-skill + ctx-to-kg.py | placeholder (Stage 1.1) |
L4c source AST | indexer (tree-sitter + 5-dim ctx embed) | placeholder (Stage 1.1) |
L5 persona-template | 4D-persona | Cycle 2 |
L6 code-analysis | claude-arch-inventory | Cycle 2 |
L7 RAG | kg-rag-toolkit + sdk-agent | Cycle 3 |
L8 orchestrator | cascade-ai/mcps/eco-system | placeholder (Stage 1.1) |
L9 multi-project | hive-mind | DEFERRED |
License
Apache-2.0. See LICENSE.
This server cannot be deployed
Maintenance
ActivitySlowing
ResponsivenessNo issues
Related MCP Connectors
Governed data discovery, exact queries, decisions, simulations, and runtime utilities over MCP.
Knowledge coverage map and health score. Ingest docs into a governed knowledge graph via MCP.
Make your knowledge agent-ready. One MCP endpoint, 5 connectors, 3 search modes.
- LensHubOAuthai.lenshub
Team knowledge from 20 connectors, served to any MCP agent — classified, scored, access-controlled.
Related MCP Servers
- AlicenseAqualityAmaintenanceEnables AI assistants to search, analyze, and understand multi-language codebases by providing indexed code intelligence via MCP.16700 npm8MIT
- FlicenseNot gradedqualityCmaintenanceEnables local RAG orchestration with MCP, providing context retrieval, in-memory text ingestion, explicit tool invocation, and transport-neutral tool discovery.-
- FlicenseNot gradedqualityBmaintenanceEnables retrieval-augmented question answering with source-grounded citations, document retrieval, and idempotent ingestion through MCP tools.-