CodeFusion Studio (CFS) is Analog Devices’ award-winning embedded software development ecosystem, built on top of Microsoft Visual Studio Code. It was designed to address the growing complexity of modern embedded engineering — where heterogeneous multi-core systems, AI-enabled workloads, and shrinking development timelines demand tools that go well beyond what legacy single-core, single-vendor IDEs were built to handle. This page introduces what CFS is, the philosophy behind it, and how its major subsystems fit together.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/analogdevicesinc/codefusion-studio/llms.txt
Use this file to discover all available pages before exploring further.
Why Analog Devices Built CFS
Embedded software engineering has undergone a fundamental shift. Engineers now configure and debug multiple processing architectures simultaneously, manage diverse toolchains and SDKs, and are increasingly expected to deploy AI models to edge hardware — all while working under tighter time-to-market pressure. Traditional proprietary IDEs were not designed for this reality: they generate inflexible code, lock developers into single-vendor ecosystems, and often become obsolete before the hardware they target reaches end-of-life. CFS was created to solve these problems with a different philosophy: open source, extensible, and designed for the long term. Every critical action that can be performed in the GUI can also be executed from the command line, enabling reproducible CI/CD pipelines and agentic AI automation. Configuration and design decisions are stored in structured, version-controllable JSON files rather than opaque binary project formats — giving teams full ownership of their workflows.Architecture Overview
CFS is composed of three integrated layers that work together:- VS Code Extension — The primary IDE surface, providing the Workspace Creation Wizard, System Planner, AI Tools panel, multi-core debugger, and all graphical interfaces. It is published on the Visual Studio Code Marketplace and auto-updates alongside the SDK.
cfsutilCLI — A command-line utility that exposes broad feature parity with graphical workflows. Use it to create workspaces, manage packages, run AI model workflows, build and flash firmware, and integrate CFS into automated pipelines — all without opening the IDE.- Plugin System — An extensible, platform-agnostic plugin architecture that separates configuration capture from code generation. Each supported firmware platform (Zephyr, MSDK, SHARC-FX) exposes its own plugin, and developers can add custom plugins to target specific RTOS, middleware, or build standards.
Key Pillars
Open Source
Apache-licensed tooling and open-source toolchains you own, modify, and integrate into your own workflows. Built-in support for Zephyr RTOS. CFS is designed for long-term use — the tools outlast any single product generation.
AI at the Edge
Complete model-to-deployment workflow: create AI-ready workspaces from model files, validate compatibility, profile runtime performance on target hardware, and generate optimized inference-ready code. The AI Debug Assistant connects GitHub Copilot or any MCP-compatible client directly to live debug sessions.
Security (TESA)
Deep integration with Analog Devices’ Trusted Edge Security Architecture (TESA), providing Secure Boot, Secure Channel, Lifecycle Management, Secure Storage, and Attestation. Arm® TrustZone® templates and Trusted Firmware-M (TF-M) support are available for qualifying devices.
Flexibility & Visibility
The System Planner provides a graphical interface for allocating memory, peripherals, clocks, and pins across multiple cores. Heterogeneous multi-core debugging runs in a single unified environment — one IDE, one debug session, one hardware debugger.
Supported Architectures and Processors
CFS supports Arm® Cortex-M, RISC-V, and SHARC-FX architectures across a wide range of ADI products. The table below summarizes capability coverage per processor family.MSDK Processors (MAX32xxx / MAX78xxx)
| Processor | Bare-metal SDK | Zephyr | Config Tools | AI Tools | TESA |
|---|---|---|---|---|---|
| MAX32650 | Yes | Yes | Yes | No | Yes |
| MAX32655 | Yes | Yes | Yes | No | Partial |
| MAX32657 | No | Yes | Yes | Yes | Yes |
| MAX32658 | No | Yes | Yes | No | Yes† |
| MAX32660 | Yes | Yes | Yes | No | No |
| MAX32662 | Yes | Yes | Yes | No | Partial |
| MAX32666 | Yes | Yes | Yes | No | No |
| MAX32670 | Yes | Yes | Yes | No | Yes |
| MAX32672 | Yes | Yes | Yes | No | Partial |
| MAX32675C | Yes | Yes | Yes | No | No |
| MAX32690 | Yes | Yes | Yes | Yes | Yes |
| MAX78000 | Yes | Yes | Yes | No | Partial |
| MAX78002 | Yes | Yes | Yes | Yes | No |
SHARC-FX Processors (ADSP-2183x / ADSP-SC83x)
| Processor | Bare-metal SDK | Zephyr | Config Tools | AI Tools | TESA |
|---|---|---|---|---|---|
| ADSP-21834 / 21834W | Yes | No | Partial | Yes | No |
| ADSP-21835 / 21835W | Yes | No | Partial | Yes | No |
| ADSP-21836 / 21836W | Yes | No | Partial | Yes | No |
| ADSP-21837 / 21837W | Yes | No | Partial | Yes | No |
| ADSP-SC834 / SC834W | Yes | No | Partial | Yes | No |
| ADSP-SC835 / SC835W | Yes | No | Partial | Yes | No |
SHARC-FX W variants (automotive-grade) are listed as separate plugin entries in CFS but have the same support level as their corresponding base processors. Config Tools support for SHARC-FX is limited to Pin Mux configuration only.
Feature Highlights
Workspace Creation Wizard
Start single- or multi-core projects in seconds with guided setup, example templates, and optional TrustZone® secure/non-secure partitioning. Create AI-ready workspaces directly from a model file with automatic compatibility checking.
System Planner
Visually configure memory partitions, peripherals, clocks, pins, and AI model assignments per core. Generates ready-to-build source code in one step, and stores all configuration in structured JSON for version control.
Multi-core Debugging
Debug multiple cores in one unified environment with breakpoints, disassembly, cross-core support, and RTOS thread awareness. Use the GDB Toolbox, Memory Viewer, Core Dump Analysis, and ELF File Explorer for deep system inspection.
AI Debug Assistant
Connect any MCP-compatible AI client to live debug sessions. The assistant can investigate hard faults, decode fault registers, inspect hardware state across cores, and trace memory corruption autonomously.
Package Manager
Download SDKs, toolchains, and plugins on demand. No full reinstall needed for updates — install only what you need, when you need it. Supports semantic versioning ranges and offline local cache installs.
CLI-First Workflows
cfsutil provides broad parity with graphical workflows for CI/CD integration, automation scripts, and agentic AI pipelines. Build, flash, configure workspaces, and manage packages without ever opening the IDE.Telemetry
CFS collects anonymous diagnostics and usage data only if you explicitly opt in. Telemetry is disabled by default. On first launch you will be prompted to choose whether to share anonymous data. You can update this preference at any time by searching forcfs.telemetry in VS Code Settings. The cfs.telemetry.enable setting also respects the global VS Code telemetry.telemetryLevel setting — if the global setting is disabled, no CFS telemetry is sent regardless of the CFS-specific setting.
Explore the Documentation
Installation
Install CFS 2.2.0 on Windows, macOS, or Linux. Covers software requirements, the setup wizard, VS Code extension install, and optional Olimex JTAG driver setup.
Setup
Configure your SDK path in VS Code settings, access the CFS Terminal, and verify your environment is ready to build and debug.
Package Manager
Install and manage SDKs, toolchains, and plugins on demand via the Command Palette or
cfsutil pkg commands.ADI Developer Resources
Full user guide, release notes, and additional resources on the Analog Devices developer portal.