Introducing QECTOR Decoder Workbench v3.5.0

Devlog: Introducing QECTOR Decoder Workbench v3.5.0

The Ultimate Cross-Compatible QEC Evaluation Platform & 29-Tool MCP Server

The infrastructure gaps in Quantum Error Correction (QEC) research are officially closed. Today, we are releasing the QECTOR Decoder Workbench v3.5.0. This desktop analysis suite functions as a complete simulation, visualization, and developer environment built on top of the high-performance qector_decoder_v3 compiled Rust core.

This release bridges raw command-line scripting with an interactive dashboard, while introducing native ecosystem integrations designed for local machines, high-performance compute clusters, and agentic AI developer setups.

🖥️ Deep-Dive: The 6 Core GUI Tabs

The frontend architecture uses a clean design featuring Consolas for code layouts and Segoe UI for telemetry readability. The layout splits your entire simulation workflow into six specific modular tabs:

• Code Explorer: Build fault-tolerant codes across 6 distinct families (Repetition, Ring, Rotated Surface, Unrotated Surface, Toric, Heavy Hex) with dynamically configurable code distances. Instantly calculate and view physical properties including qubit footprints, check nodes, code distances, and code rates.
• Decoder Lab: An interactive, single-syndrome sandbox featuring 5 distinct, cross-compatible decoders (union_find, fast_union_find, blossom, sparse_blossom, and bp_osd). Tune physical error rates and seed configurations inline to watch error chains evolve.
• Benchmark Suite: Configure thorough execution loops and pull high-precision latency statistics including Mean, P50, P99, Min, and Max execution boundaries. Export raw benchmark metrics natively to clean JSON tables.
• Batch & Streaming: Run bulk data sets with explicit CPU/CUDA/OpenCL hardware routing. The engine eliminates silent processing fallbacks—if a high-performance backend is unavailable, it fails loudly so you can diagnose configuration issues instantly. Features sliding-window FIFO commit semantics for streaming decodes.
• Hardware & System: Auto-detects local host configurations (cuda_is_available(), opencl_is_available()). The interface features a built-in recommendations module (hardware_routing.recommend()) to match your current system specs with the optimal backend decoder routing configurations.
• Documentation Studio: Export complex simulation data, provenance metadata, and multi-algorithm comparison tables across multiple layouts including Markdown, JSON, HTML, LaTeX, PDF (via matplotlib PdfPages), SVG, and vector images.

🤖 The 29-Tool Model Context Protocol (MCP) Server

To support modern, AI-assisted development pipelines, Workbench v3.5.0 ships with a built-in 29-Tool Model Context Protocol (MCP) Server.

This isn't a slow HTTP bridge or an open local network port. It implements a clean, newline-delimited JSON-RPC 2.0 transport architecture directly over standard I/O (stdin/stdout) following the modern 2024-11-05 protocol specification. Every single one of the 29 tools is linked directly to the underlying backend.py API. This allows local LLMs and agentic developer setups to call your native compiled decoders, map Tanner graphs, and execute syndrome corrections automatically via standard tooling prompts.

[ LLM IDE / Cursor / Claude ] ──► (stdin/stdout JSON-RPC 2.0) ──► [ 29-Tool MCP Server ] ──► [ Rust Backend ] 

⚖️ Licensing Architecture & The Open PyPI Boundary

We are maintaining an honest, transparent posture regarding where this app sits in your stack. The software is split into two clear licensing domains:

1. The Workbench App (This UI Tool): Distributed under a highly permissive source-available EULA. It grants a royalty-free, worldwide license to copy, execute, use, and distribute the app for any purpose—including academic, personal, and full corporate use—as long as embedded watermarks and "QECTOR" notices remain intact.
2. The Backend (qector-decoder-v3 library): A separately licensed platform. While the Workbench UI layer can be distributed freely, the underlying compiled decoding core running behind it is free only for personal, educational, and academic research. Commercial operations, enterprise R&D teams, or automated hardware tuning loops utilizing the core must maintain a paid commercial license via qector.store.

🚀 Get Started

The suite includes an integrated boot-time auto-updater that queries PyPI to ensure your underlying runtime engines match the newest stable qector_decoder_v3 releases automatically.

• Download the Windows Portable Executable: Grab QectorWorkbench-V3.5.0-Portable.exe directly from our itch.io repository.
• Corporate Clearance: Ensure your team clears corporate compliance alerts by grabbing a commercial evaluation token at qector.store/pricing.

Let me know if you would like me to draft an accompanying social media teaser for X and LinkedIn to announce the itch.io launch, or if we should refine the JSON-RPC schema documentation for the 29 MCP tools!

https://id01t.itch.io/qector-decoder-workbench