Frequently Asked Questions

Zoo Design Studio is a robust, unified CAD platform built on our proprietary geometry engine, combining AI-driven generation and modification of boundary-representation CAD files with the traditional sketch and feature tree workflows engineers already know, and includes Zookeeper, our conversational CAD agent.

Zoo Design Studio is best used by downloading and installing our native desktop applications. There is a web browser test environment but it is meant to be used as more of a testing environment.

Visit our downloads page to install the desktop application.

Zoo Design Studio is available for free, and our free version includes 20 minutes of free Zookeeper reasoning time. We have paid tiers that include additional Zookeeper reasoning time. You can view our pricing here.

No. Zoo Design Studio is CAD software that anyone can use and download for free. There are paid subscription tiers that unlock more powerful features and capabilities, like unlimited Zookeeper credits and support for teams and enterprises. Visit our pricing page for more details.

Yes. Zoo Design Studio requires an internet connection because processing and our geometry engine run in the cloud. This lets you handle complex CAD workflows smoothly, even on low-powered machines.

Zoo Design Studio has solid modeling features like extrude, revolve, sweep, booleans, etc. For a full list of features, visit our documentation.

Yes. Assemblies have been supported in Zoo Design Studio since v1.0.

We constantly ship updates to Zoo Design Studio. We do not currently follow a scheduled release cadence—this allows us to release exciting features and bug fixes as quickly as we can.

These are the current file types that are compatible with Zoo Design Studio:

• STEP/STP
• SLDPRT
• KCL
• STL
• glTF/GLB
• OBJ
• FBX
• PLY

You can export your project into the following file types:

• STEP
• STL
• glTF
• OBJ
• PLY
• FBX

Zookeeper lets engineers delegate tasks such as research, reasoning, and manufacturing-aware design decisions to an AI agent that designs based on their goals and descriptions. It conducts relevant research, proposes design constraints and parameters, and generates a clear design plan before producing a CAD model. The resulting model is fully editable using the same sketch and feature tree workflows engineers already rely on. Building on our Text-to-CAD experience, the team added research and reasoning capabilities and engine-level tools that allow the agent to inspect, snapshot, and debug geometry while generating production-ready CAD.

Anyone from beginners and non-engineers to expert CAD users and enterprise teams can benefit from Zookeeper. It speeds up design by turning plain text requests into editable parametric KCL models for newcomers, while saving time for more experienced users with rapid prototyping, project-aware edits, and repeatable workflows that scale for teams.

Zookeeper is a pane on the right side. Click the sparkles icon in the right control bar to open it. It's context-aware and works with your active project to create or edit designs.

Zookeeper can generate a wide range of CAD models, from simple primitives and brackets to mechanical parts, enclosures, mounts, assemblies, and 3D-printable components, all from prompts. It can also edit existing models, interpret selected geometry and renders, and update the active project by producing or modifying parametric KCL code.

Yes - Describe what you want and Zookeeper will try to generate the geometry as parametric KCL. If generation fails or needs tweaking, it will suggest prompt refinements or follow-up edits.

Yes - Zookeeper analyzes your active project and offers context-aware suggestions (e.g., add fillets, mounting holes, or adjust dimensions) to help you decide the next modeling steps. It can also propose small edits or refinements you can apply immediately. It conducts relevant research, proposes design constraints and parameters, and generates a clear design plan before producing a CAD model.

Yes - Zookeeper understands and uses your selections in the active project. When you select faces, edges, bodies, or code regions, Zookeeper uses that context to make targeted edits, add features, or generate more precise prompts.

Yes - Zookeeper can view and interpret the project's renders and multiview snapshots to understand geometry and suggest or make context-aware edits. It uses those visual cues alongside selections and KCL to produce more accurate modifications. The team added engine-level tools that allow the agent to inspect, snapshot, and debug geometry while generating production-ready CAD.

Yes - You can ask Zookeeper questions about modeling, KCL, design decisions, or your active project. It can research using the internet, explain tradeoffs, and make manufacturing-aware suggestions. When you ask, Zookeeper can also perform edits, apply changes, or generate models directly in your design.

Yes - Zookeeper can compute surface area, volume, mass, and center of mass for the active project and return results in common units (e.g., mm³, m², kg). It uses the model geometry and material/density info from the project to make these calculations.

Be specific about what you want (dimensions, constraints, intended use) and include selections or reference geometry in the active project - clear prompts produce more accurate KCL results. Iterate: review the generated model, tweak parameters or ask for targeted edits, and accept Zookeeper's suggestions (fillets, holes, thickness changes) to quickly refine the design.

Zookeeper understands natural-language prompts (primarily English) and can read and work with KCL code and selected geometry in your project. It also accepts many major languages (Spanish, French, German, Chinese, etc.) with varying accuracy, though the best results are typically in English.

KCL (KittyCAD Language) is Zoo's own scripting language for describing CAD geometries and rendering them within Zoo Design Studio. Every CAD function in Zoo Design Studio generates KCL. Using KCL in Zoo Design Studio is completely optional and is not required. KCL allows designers to leverage the benefits of code in the CAD ecosystem. Learn more by reading our KCL book.

No! Coding with KCL (Zoo's CAD coding language) is completely optional. However, it can be quite powerful. If your curiosity gets the best of you, you can learn more about getting started with KCL for coding in CAD Community Page.

Yes, the Zoo Design Studio application is open source and you can find it on GitHub. However, our CAD geometry engine is not open source.

Check out our developer documentation to get started designing your own modeling tools.

For assistance, check out our Community Page! See if others have asked for help with the same thing and if not, make a new topic so we can help you. Users who are on a paid plan can use email support or submit a support ticket here if they are logged in.

There are a few places you can report a bug. You can report it on our Community Page, you can let us know on our Discord, or if you are familiar with GitHub, you can make an issue for us.

If an update is needed, a notification will appear the next time you open Zoo Design Studio. Once the update is complete, you will need to restart your application for it to take effect.

You can use our Community Page to see if someone else has requested the same thing. If no one has, make a new request using the feature category.

Zookeeper lets engineers delegate tasks such as research, reasoning, and manufacturing-aware design decisions to an AI agent that designs based on their goals and descriptions. It conducts relevant research, proposes design constraints and parameters, and generates a clear design plan before producing a CAD model. The resulting model is fully editable using the same sketch and feature tree workflows engineers already rely on. Building on our Text-to-CAD experience, the team added research and reasoning capabilities and engine-level tools that allow the agent to inspect, snapshot, and debug geometry while generating production-ready CAD.

Anyone from beginners and non-engineers to expert CAD users and enterprise teams can benefit from Zookeeper. It speeds up design by turning plain text requests into editable parametric KCL models for newcomers, while saving experts time with rapid prototyping, project-aware edits, and repeatable workflows that scale for teams.

Zookeeper is a pane on the right side. Click the sparkles icon in the right control bar to open it. It's context-aware and works with your active project to create or edit designs.

Zookeeper can generate a wide range of CAD models, from simple primitives and brackets to mechanical parts, enclosures, mounts, assemblies, and 3D-printable components, all from prompts. It can also edit existing models, interpret selected geometry and renders, and update the active project by producing or modifying parametric KCL code.

Yes - Describe what you want and Zookeeper will try to generate the geometry as parametric KCL. If generation fails or needs tweaking, it will suggest prompt refinements or follow-up edits.

Yes - Zookeeper analyzes your active project and offers context-aware suggestions (e.g., add fillets, mounting holes, or adjust dimensions) to help you decide the next modeling steps. It can also propose small edits or refinements you can apply immediately. It conducts relevant research, proposes design constraints and parameters, and generates a clear design plan before producing a CAD model.

Yes - Zookeeper understands and uses your selections in the active project. When you select faces, edges, bodies, or code regions, Zookeeper uses that context to make targeted edits, add features, or generate more precise prompts.

Yes - Zookeeper can view and interpret the project's renders and multiview snapshots to understand geometry and suggest or make context-aware edits. It uses those visual cues alongside selections and KCL to produce more accurate modifications. The team added engine-level tools that allow the agent to inspect, snapshot, and debug geometry while generating production-ready CAD.

Yes - You can ask Zookeeper questions about modeling, KCL, design decisions, or your active project. It can research using the internet, explain tradeoffs, and make manufacturing-aware suggestions. When you ask, Zookeeper can also perform edits, apply changes, or generate models directly in your design.

Yes - Zookeeper can compute surface area, volume, mass, and center of mass for the active project and return results in common units (e.g., mm³, m², kg). It uses the model geometry and material/density info from the project to make these calculations.

Be specific about what you want (dimensions, constraints, intended use) and include selections or reference geometry in the active project - clear prompts produce more accurate KCL results. Iterate: review the generated model, tweak parameters or ask for targeted edits, and accept Zookeeper's suggestions (fillets, holes, thickness changes) to quickly refine the design.

Zookeeper understands natural-language prompts (primarily English) and can read and work with KCL code and selected geometry in your project. It also accepts many major languages (Spanish, French, German, Chinese, etc.) with varying accuracy, though the best results are typically in English.

KCL (KittyCAD Language) is Zoo Design Studio's domain-specific, parametric CAD programming language. KCL projects are collections of .kcl files (entrypoint: main.kcl) that programmatically create 3D geometry using primitives, sketches, transforms, booleans, and parameters.

Zookeeper is our built-in CAD design assistant inside Zoo Design Studio that helps you work with KCL: it can generate new KCL models from plain-language prompts, edit your existing .kcl code, answer questions about KCL, and propose design changes, acting as an intelligent, context-aware partner while you design.

Yes - Zookeeper can read your KCL code, explain syntax and structure, and help fix or improve scripts. It can also suggest idiomatic KCL patterns and make project-aware edits when you ask.

Yes - Zookeeper can read and understand your KCL code, explain what it does, and use that context to suggest fixes or make targeted edits to the active project.

Yes - Zookeeper references official KCL materials (including the KCL Book) to ensure advice and code follow best practices. See the KCL Book for details.

Yes - Zookeeper can identify and fix KCL errors by analyzing your project and selected code, then applying targeted edits. If an automatic fix can't be applied, it will explain the issue and suggest a clear repair or simplified prompt for you to approve.

Yes - Zookeeper can explain mechanical design concepts, recommend best practices (tolerances, fastening, draft, fillets), and help apply them to your model. It can provide guidance and simple calculations, but it doesn't replace engineering validation or guarantee manufacturability or structural performance.

• Can: create/edit KCL CAD models, explain KCL/design concepts, and make changes via the studio's AI actions.

• Cannot: run physical tests, guarantee manufacturability or structural safety, or perform detailed engineering simulations.

• Interaction rules: modifies projects only through the studio's built-in code actions and cannot paste raw KCL or show rendered images in chat.

• Knowledge bounds: limited to its training and available docs; ask for clarification when requests are vague.

• Change ZDS application settings.

• Control the viewport (you must rotate, pan, zoom manually).

• Open or switch to other projects.

• Export models or files.

• Ingest arbitrary images or external 3D files as input.

• Display backend-generated progress images.

• Read files outside the active project directory.

• Access your Zoo account or personal account information.

Zookeeper makes an initial attempt and, if it encounters errors, will examine the error messages and try up to five automated fixes. If those attempts still fail, it stops, reports the problem, and offers simpler prompts or clarification options for you to choose.

If a Zookeeper prompt fails, it stops and shows the error(s), explains what went wrong, and offers next steps - like a simpler prompt, clarification questions, or manual edits. It may try up to five automated fixes for edit actions; if those still fail it won't retry automatically and will wait for your instruction to proceed.

Yes. Zookeeper can make mistakes - it may misunderstand intent, produce incorrect geometry, or suggest designs that aren't manufacturable or safe. Always review and validate changes, run your own checks, and give clear, specific instructions or corrections when needed.

Short answer: Not persistently. You can give explicit instructions, constraints, or coding style rules in each prompt (and selections/params in the project guide edits), and Zookeeper will follow them for that action. It doesn't support global, always-on custom agent rules or persistent personality settings across sessions.

Break the task into smaller steps. Ask Zookeeper for a design plan first, then request geometry changes one feature at a time. Example flow:

1. "Create the rectangular base."

2. "Add corner holes."

3. "Fillet the edges."

Working incrementally (and giving clear dimensions or constraints) gets better results than one large, vague prompt.

Select the exact geometry and give a concise instruction.

• In the UI, highlight the face/edge/body or select the relevant lines in the code before asking.

• Use clear, specific commands (e.g., "change this hole diameter to 5 mm" or "fillet these two edges with 2 mm").

• Reference named parts or parameters if available (e.g., "modify mount_plate.thickness").

• Provide constraints or intent when helpful (fit a M4 screw, keep wall ≥2 mm).

Doing this ensures Zookeeper applies changes to the correct part rather than guessing.

You can report bugs many ways.

1. Submit a Github issue

2. Report bugs on our Discord server

3. Ask our team via our Community Forum

Short answer: Only in limited ways for now. Zookeeper's API features are narrow but expanding. Currently you can use the Text-to-CAD endpoint to generate KCL models from prompts; more API-powered capabilities will be added over time.

Zookeeper usage is included with Zoo Design Studio plans as monthly reasoning time.

• Free plans include 20 minutes per month

• Plus plans include ~400 minutes per month

• Pro plans include unlimited reasoning time

If you use up the reasoning time included in your plan, you can continue using Zookeeper on a pay-as-you-go basis at $0.0083 per second, or upgrade to a higher plan.

The Free plan gives you full access to Zoo Design Studio, including all core CAD workflows. Zookeeper is included with 20 minutes of reasoning time per month, and Zoo may train on data from Free plan usage to improve our products.

We'll notify you when you're close to or have used up your included reasoning time. You can upgrade at any time from your account settings. If you run out of included reasoning time, you can either continue using Zookeeper on a pay-as-you-go basis at $0.0083 per second or upgrade to a higher plan with more included reasoning time.

We do! You can pay month-to-month and contact us to cancel any time. If you opt for an annual plan, you get a 20% discount in exchange for paying upfront and remaining committed to that plan for an entire year.

We accept credit cards for all paid plans. For Team or Enterprise plans, if you need custom billing, please contact us.

Yes, you can always upgrade or downgrade your plan to another tier without losing your data.

Visit our Getting Started documentation and YouTube channel for the latest onboarding resources.

On the Free Plan, the Zookeeper reasoning time is capped at 20 minutes per month. It resets on the first of every month. If you go over in a given month, you won't be able to use Zookeeper again until the first day of the following month, or you can upgrade to any of our paid plans and get unlimited Zookeeper balance!

Please contact us to discuss pricing and features for your company. We will work with you to craft a plan that optimizes value and works best to fit your goals.

If you're an enterprise user, we'd love to learn more about your custom workflows. Please contact us so we can discuss.

Yes! This is extremely powerful and can provide significant value for organizations with large CAD operations. Contact our team to learn more about how we can help.

Zoo is in the process of obtaining SOC2 Type II certification (expected by Q4 2025). Zoo adheres to industry standards for safeguarding your data.

Zoo offers a full suite of APIs that can connect to any platform. Please refer to our developer documentation for more information.

API calls are charged by the second. Each second costs ~$0.0083.

Rounding: We compute elapsed_seconds x $1/120, then round Half-Up to the nearest cent per call. We never round the time itself.

Short calls policy: Historically, sub-10-second calls were not billed. Starting October 14, 2025, we will bill them per second with the same rate and rounding.

Note: because we round each call to cents, many tiny calls can sum to a hair more than one long call with the same total seconds. This is standard for usage billing. If that ever meaningfully impacts you, tell us.

For more information on how the math works, see Jess's blog post here.

All users get $10.00 worth of API calls per month for free. After that, it is pay-as-you-go pricing.

Yes, all of our charged API calls are charged by the second. Each second costs ~$0.0083. For more information on how the math works, see Jess's blog post here.

No, you are not charged if the API call fails.

Generally no. GET requests are free, except for WebSocket-style GETs used for real-time features: endpoints that start with /ws/modeling/commands or contain /ws/ml/copilot are billable.

In general, any POST endpoint is billable unless it is on our exclusion list below. The only GET endpoints that are billable are WebSocket-style GETs for real-time features: endpoints that start with /ws/modeling/commands or contain /ws/ml/copilot.

Based on our current API specification (spec.json), representative billable POST endpoints include:

• /ai/text-to-cad/{output_format}

• /ml/kcl/completions

• /ml/text-to-cad/iteration

• /ml/text-to-cad/multi-file/iteration

• /ml/convert/proprietary-to-kcl

• /file/conversion

• /file/conversion/{src_format}/{output_format}

• /file/mass

• /file/volume

• /file/surface-area

• /file/density

• /file/center-of-mass

This list is not exhaustive; new billable endpoints may be added in the future. Always refer to the API spec and this FAQ for the latest billing behavior.

No. Endpoints that start with /user or /org are free and not billed.

No. Endpoints that start with /oauth2 are free and not billed.

The following endpoints are free and never billed:

• /_meta
• /user
• /logout
• /apps
• /events
• /org
• /store
• /_internal
• /auth
• /oauth2
• /debug
• Any endpoint containing shortlinks
• /file/execute/
• /ws/ml/reasoning/{id}

Text-to-CAD generations can vary in their duration. It depends on the complexity of your prompt. A large majority of Text-to-CAD API calls last between 10 - 30 seconds. However, they can last minutes if the prompt is long and complex.

Yes. We are working on our machine learning endpoints every day, but it is still experimental. Some results may not be as good as you expect.

Text-to-CAD includes a STEP file and can optionally output KCL as well with every call. For any other file format, you will be charged for the file conversion API used.

For assistance with APIs, you can ask a question on our Community Page or discuss on our Discord!

For bugs, mention it to us on our Community Page.

Please email us directly. We would love to chat with your specific needs and see how we can best help you.