To engineer boards or mechanical parts using Codex, your engineers must connect external software, develop custom scripts, and purchase individual tool licenses. Every engineer operates an isolated agent, then the team merges files via git. Within Flux, the necessary hardware tools are built in. Your engineers and the AI engineers collaborate simultaneously inside a single live project.
Flux operates as an autonomous hardware engineer for your team. Give it a requirement. It organizes the project, selects the parts, draws the schematics, routes the board, creates mechanical parts, and codes the firmware. It verifies its results using simulation and design rule checks. Your engineers receive manufacturing files.
Codex is the coding agent from OpenAI. It is included in ChatGPT Free, Go, Plus, Pro, Business, Edu, and Enterprise plans.[1]
Codex works on coding tasks in the cloud, from bug investigations to new features. Each task has its own workspace. You review the changes and open a pull request when you are ready.[2]
Each Codex task runs in its own cloud workspace and continues while your computer is asleep.[2]
You inspect the changed files and test results, ask for follow-up changes, and open a pull request.[2]
Codex comes with eligible ChatGPT plans, on the web, on mobile, and in the desktop app.[2]
Codex can write scripts for hardware tasks. Then you pick the outside tools, write the code that links them, check the output, and keep the whole setup running. That takes time and expertise, and each script adds one more part that can fail. Every outside tool also needs a separate license or account. To connect to Fusion 360, for example, you need a Fusion 360 license. Flux puts the same frontier models in a complete hardware environment: a CAD and EDA editor, tools to inspect the design, 734,000+ public parts with live data from distributors, simulation and design rule checks, import from common CAD and EDA formats, manufacturing files, and one shared live project for the team. Everything is built in, tested, part of your Flux plan, and ready from the first prompt. Codex runs OpenAI models. Flux is model agnostic, so no single model provider locks you in. Flux works in your web browser. You install nothing, and your projects are stored in the cloud.
| Feature | Flux | Codex |
|---|---|---|
| Requirements and specification | The AI engineer turns your idea into requirements and a product specification, and checks the design against them. | Writes documents and code from your prompt.[6] |
| What it is | AI hardware engineer with internal design software | General AI coding agent[1] |
| Where it runs | Works inside your web browser. No installation required. Your projects stay in the cloud. | Tasks run in cloud workspaces. You use it from the web, the desktop app, mobile, or the command line.[2] |
| Who does the design work | The Flux agent. Your engineers review and edit. | Developers assign software tasks to the coding agent.[6] |
| Schematic and parts | Included. Selects from 734,000+ public parts. | Not built in. You connect outside tools and part data with scripts.[1] |
| Board layout and routing | Integrated | Not built in. You connect an outside board tool with scripts.[1] |
| Placement and routing | The AI engineer places the parts and routes the board like an experienced designer. Your engineers review and edit. | No built-in PCB trace-routing tools.[6] |
| High-speed design | The AI engineer plans DDR topology and timing budgets, routes differential pairs and length-matched buses, tunes length and skew, fans out BGAs with blind, buried, and microvias, and sets trace impedance from your stackup. It checks the routed board with S-parameters, TDR, eye diagrams, crosstalk analysis, and a signal integrity, power integrity, and EMC review. Your engineers review and edit. | No built-in PCB signal-integrity analysis.[6] |
| Parts and sourcing | 734,000+ public parts, with live prices and stock from distributors | No built-in electronics parts library.[6] |
| Simulation and design rule checks | SPICE and power delivery network simulation, full-wave electromagnetic and signal integrity simulation, thermal simulation, and mechanical simulation; Flux runs design rule checks. | Not built in. You connect outside simulators with scripts.[1] |
| Firmware | Codes firmware for the board it created | Writes code, including firmware[2] |
| Wiring harnesses | Built in. Flux designs the wiring harnesses between the boards: connectors, pinouts, and wires, in the same project. | No built-in wire-harness design tools.[6] |
| Manufacturing files | Gerber, ODB++, IPC-2581, bill of materials, pick-and-place, and more. Flux can check the files before you order. It checks copper gaps, silkscreen, and board outline, then checks the BOM and pick-and-place files. | Not built in. Your scripts or outside tools write the files.[1] |
| Quotes and orders | Quotes and orders for boards, 3D printing, CNC machining, and wire harnesses, from the project. | No built-in hardware quoting or ordering workflow.[6] |
| Import from other tools | Imports complete projects from Altium, OrCAD and Allegro, KiCad, PADS and Xpedition, CADSTAR, Eagle, Fusion Electronics, EasyEDA, DipTrace, and SOLIDWORKS PCB. Flux imports Labcenter Proteus SDF netlists with Specctra board exports. IPC-2581 revision B and C files become board projects with reusable footprint parts. IDF 2.0 and 3.0 board files become board projects with the outline, cutouts, drilled and mounting holes, part placements, and 3D part heights. | No built-in EDA project importer.[6] |
| Versions, branches, and merges | Full change history. The AI engineer forks a design into a second version, keeps both, and merges the changes that you pick. | No native hardware design version history.[6] |
| Who reviews the work | Your entire team shares a single live project in the browser: electrical, layout, mechanical, and firmware engineers | Developers review the code changes[2] |
| Product images, videos, and documents | Product images, product videos, datasheets, and regulatory documents, made from the design in the project. | Writes product documentation and code from your prompt.[6] |
| Mechanical CAD | Integrated and kernel agnostic. Ranging from an enclosure to a complete robot arm. | Not built in. You connect an outside CAD kernel with scripts.[1] |
| Mechanical simulation | Finite-element stress, thermal, and airflow simulation | No mechanical simulation workspace.[6] |
| Mechanical manufacturing | STEP, STL, and 3MF files, plus quotes for 3D printing, CNC machining, sheet metal, and laser cutting | No CNC manufacturing workspace.[6] |
| CNC toolpaths and G-code | The AI engineer writes the toolpaths, the G-code, a toolpath preview, and a setup sheet for the parts that it designed. | No CNC toolpath or G-code workflow.[6] |
| AI models | Frontier models from multiple providers, including Claude and OpenAI models | OpenAI models[1] |
| Tool licenses | Your Flux subscription contains the editor, mechanical CAD, parts data, simulation, and manufacturing files. Zero extra purchases required. | Each outside tool needs its own license or account, for example a Fusion 360 license[5] |
| Team knowledge and rules | Document your standards once for the organization. Each AI engineer applies these guidelines on all projects. Flux checks all designs against your rules. | AGENTS.md files in each code repository[3] |
| Parallel work | A primary AI engineer separates a product into tasks across multiple projects. It runs them simultaneously and provides status updates. Major changes require your approval. | Many coding tasks in the cloud at the same time[2] |
Start a Flux project free and turn your hardware requirements into a complete device.
Orchestration
Provide a complete product to Flux. A lead AI engineer divides the work into tasks, initiates parallel operations across projects, manages them, and reports back.
Every task operates inside a dedicated agent process. The main board, the daughterboards, and the enclosure progress concurrently.
The lead engineer starts tasks in other projects and tracks everything from one location.
It monitors every active task, corrects work that goes off target, and collects the results.
Modifications that shift an interface or change board size wait for your approval.
Flux drafts a specification for each board covering its outline, mounting holes, connectors, and the corresponding enclosure cutouts. Every board receives an independent project and a dedicated AI engineer for parallel development. Flux then inserts the completed boards into the main assembly to measure clearances and connector alignment in millimeters.
Built for teams
Flux operates inside the web browser. Your entire team accesses one live project: engineers, reviewers, and AI agents working together. No local CAD installation is necessary.
Each design, revision, and comment resides inside a single cloud workspace accessible to your entire team.
Libraries, design rules, and past decisions stay with the team. The AI starts every job completely informed.
Initiate AI tasks, verify the results, and continue engineering where a teammate stopped.
Open a link and start engineering. You have no licenses to manage and no workstations to configure.
Speak to your AI engineer instead of typing, using your computer or your phone.
Customizable
Every team builds hardware differently. Provide your standards to Flux, connect your tools, and let the AI build the small applications each project requires.
Onboard Flux the way you onboard a new engineer. Write down your design guidelines, preferred parts, and past lessons once. Every AI engineer in your organization applies this knowledge on every project.
Define your layout rules and fabricator constraints once. Flux verifies every design against these parameters consistently.
Link any MCP server, or supply a web API key to Flux. Push firmware to your repository, analyze crash reports from field devices, or use the tools your team built internally.
Administrators add API keys once for the organization. The AI engineer uses them without seeing them. Flux attaches each key only to requests sent to your approved services.
Request a custom tool, and Flux generates a live page inside the project. Examples include a power budget that recalculates when you change a load, a hardware checklist tracking progress, or an interactive 3D assembly you can disassemble.
Operate Flux from Claude Code, Cursor, or any agent providing MCP support. Your external agents can read and modify Flux projects, and they can assign tasks to the Flux AI engineer.
Connects to the services your team already uses.
Onshape 
Autodesk Fusion 
PTC Windchill 
Siemens Teamcenter 
Arena PLM 
OpenBOM 
Duro 
GitHub 
GitLab 
Memfault 
Golioth 
Particle 
Nordic nRF Cloud 
Edge Impulse 
NI SystemLink 
SimScale 
Octopart 
SiliconExpert Supply a specification to Flux. The AI organizes and completes the design work independently in the cloud. The work continues after you close your computer. Your engineers inspect the final output.
Each AI engineer uses an independent Linux or Windows cloud computer containing a desktop and web browser. It operates Flux's engineering tools and the software your team installs, generates custom scripts to check its work, and keeps each setup for the next project.
A software assistant might generate scripts for schematics, simulation, and manufacturing files. You must evaluate this code, maintain it, and purchase licenses for all connected applications. Flux provides a built-in CAD and EDA editor, checking utilities, simulation, file importers, and factory outputs within your subscription. The AI begins engineering immediately.
An agent might create a simulation wrapper quickly. It cannot generate a library of 734,000+ public parts or live distributor data. Flux verifies each part using its datasheet and live distributor data. It also obtains quotes from fabricators.
Software bugs allow quick fixes and redeployments. In hardware, an incorrect footprint or skipped test requires ordering another board and delaying the project by weeks. Flux validates the design using simulation and design rule checks before your engineers review it.
Generates videos of the product it designed for team reviews, customer updates, and marketing websites.
Model-agnostic AI
Choosing between Codex and Flux is unnecessary. Flux operates as a model agnostic platform. It uses frontier models from multiple providers, including OpenAI models. Adaptive model selection assigns every engineering task to the most cost-effective model available. You achieve frontier performance at 30% lower cost.
Adaptive divides all engineering assignments into tasks and assigns every task to the most cost-effective model. Complex design decisions require frontier models. Routine operations use fast, inexpensive models. Separate tasks run concurrently, allowing complete products to finish faster. Your team interacts with a single AI engineer while Flux manages the task distribution internally.
Work with the frontier models your team already trusts, directly within Flux.
Each model receives the specialized engineering tools, project context, and workflows from Flux.
When a new model ships, your team gets it without changing engineering platforms.
Prevent depending on a single model provider for your engineering capacity.
Adaptive pairs every engineering step with the most effective model, ranging from complex design decisions to basic inspections.
One platform
Flux is not a collection of isolated CAD applications. It provides a single AI-native engineering platform where all disciplines share the same project, the same data, and the same AI engineer.
Specifications, part selection, schematic creation, and circuit simulation.
Part placement, trace routing, layer stackups, design rule checks, and cable connections between the boards.
Parametric enclosures, brackets, and mounts aligned with the actual board's mounting holes and connectors. Import your custom STEP models. A single parts library supports the entire product: chips and connectors, modules not on the board, and mechanical hardware including screws and standoffs.
Custom firmware written for the board created by Flux, targeting Arm Cortex-M, RISC-V, AVR, or Pico. The code is compiled, analyzed, and unit tested.
Real-time component sourcing, pricing, fabricator quotes for boards and cables, alongside factory-ready outputs.
Verification lab
Flux goes beyond generating designs. The AI engineer simulates and checks its own work using professional-grade solvers, and it shows you the results.
SPICE simulation of analog, power, and mixed-signal circuits, run and evaluated by the AI engineer.
Full-wave simulation of the routed board: impedance, S-parameters, crosstalk, and antennas.
Heat and airflow simulation for your boards and enclosures before you build them.
Airflow and cooling analysis inside the enclosure ensures that vents, fans, and heat sinks function properly before the first prototype.
Finite-element stress analysis for brackets, enclosures, and mounts.
Electrical and manufacturing rule checks performed on every schematic and board.
Analyzes the datasheet of every part and checks the design against its pinouts, ratings, and limits.
Sourcing
Flux connects directly with major electronics and mechanical parts distributors. The AI engineer selects in-stock chips, connectors, screws, and standoffs. It compares pricing across distributors and replaces obsolete parts before the product goes to production.
Digi-Key 
LCSC 
Mouser 
Arrow 
Verical 
TME 
HQonline 
McMaster-Carr 
MISUMI 
JLCMC Manufacturing
Flux obtains quotes directly from fabricators for circuit boards and assembly, 3D printing in plastic and metal, CNC machining, sheet metal, laser cutting, injection molding, casting, and wiring harnesses. The AI engineer compares prices and lead times, verifies the design against fabricator rules, and updates the quote as the design changes.
JLCPCB 
PCBWay 
OSH Park 
AISLER 
Craftcloud JLC 3D printing PCBWay 3D printing 
Shapeways 
OSH Cut 
SendCutSend 
MiniProto You do not have to choose. Codex connects to MCP servers, and Flux has one. Codex can read and edit Flux projects and hand design jobs to the Flux AI engineer. Your developers keep Codex for the software, and the hardware work goes to Flux.[4]
For learning and evaluation, with light AI usage and unlimited manual editing.
For active hardware projects, with included AI usage and additional usage available.
For larger projects and heavier AI usage, with hands-on support from experienced electrical engineers.
For teams on shared projects, with pooled AI usage, member seats, and priority support.
Eligible individual plans include a free trial. Teams subscriptions start paid; trial and pricing terms depend on your current offer. See all plans
Not ready to pick a plan? Start for free
Use Codex for software development inside your code repositories. Use Flux when your final deliverable must be a board, mechanical parts, and firmware prepared for manufacturing.
Review our subscription plans, or schedule a demonstration with our team.
