MCP App Store
by KeenEthics
Autodesk Flow Studio icon

Autodesk Flow Studio MCP Server for ChatGPT

By Autodesk, Inc.

Overview

Create images and 3D models from a prompt or a photo with Autodesk® Flow Studio in ChatGPT, and keep refining them in the chat. What you can do: - Create an image from a text prompt, then edit it by describing the change or generate a new variation. - Turn a prompt or an image into a draft 3D model and view it in an interactive viewer. - Refine a model by rebuilding its mesh at the level of detail you need, and retexture it from a prompt or a reference image. - Download models as GLB, OBJ, STL or 3MF for 3D printing, games, film and design work. - Browse your Flow Studio library to pick up where you left off on past projects and generations. - Pick the AI model for each step and save the ones you use most as defaults. What you need: - An Autodesk account with a Flow Studio Lite plan or higher. About Autodesk The world's designers, engineers, builders, and creators trust Autodesk to help them design and make anything. From the buildings we live and work in, to the cars we drive and the bridges we drive over. From the products we use and rely on, to the movies and games that inspire us. Autodesk's Design and Make Platform unlocks the power of data to accelerate insights and automate processes, empowering our customers with the technology to create the world around us and deliver better outcomes for their business and the planet. For more information, visit autodesk.com.

How to connect

  1. ChatGPT: Open the Autodesk Flow Studio listing in the ChatGPT app directory and choose Connect. ChatGPT listing Autodesk Flow Studio exposes 21 tools here.

Tools

check_generation_access_billing

ChatGPT
Checks whether the signed-in user's Flow Studio license allows the generation tools, and picks the license to bill. Use when: the user starts generating images or 3D models, a generation tool reported a license, credit or age-consent block, or the user says they upgraded or confirmed their age. Do not use when: the user asks about plans or remaining credits (use list_licenses_billing). Returns: allowed, outcome, message, generativeTools (per-tool access) and licenseId, the input to every credit-spending tool, valid for the rest of the chat. A blocked account is a normal result, not an error. Shows: the upgrade card for a blocked account: a message and a button to the plans page, or to Flow Studio to confirm age.

create_3d_model_from_image_mesh

ChatGPT
Creates a draft 3D model with up to four variants from an image, in a new Flow Studio mesh project. Use when: the user asks to make a 3D object from an image, including one attached in the same message, or from an image edited to add or change parts of an earlier 3D model. Do not use when: there is no image, only text (use create_3d_model_from_text_mesh). Effects: Saves an attached or linked image to the user's account and creates a new project in the user's Flow Studio library. Spends credits from the user's Flow Studio license. Returns: generationId, the input to remesh_and_retexture_draft_mesh, and sourceImageId, the image it used. Shows: the 3D viewer with progress and the finished model.

create_3d_model_from_text_mesh

ChatGPT
Creates a draft 3D model with up to four variants from a text prompt only (no source image) in a new Flow Studio mesh project. Use when: the user wants a 3D model from a text description. Do not use when: the user has an image to start from (use create_3d_model_from_image_mesh). Effects: Creates a new project in the user's Flow Studio library. Spends credits from the user's Flow Studio license. Returns: generationId, the input to remesh_and_retexture_draft_mesh. Shows: the 3D viewer with progress and the finished model.

create_image_from_text_mesh

ChatGPT
Creates images from a text prompt in a new Flow Studio mesh project. Use when: the user wants a new image from a text description, on its own or as the source for a 3D model. Do not use when: the user wants to change an existing Flow Studio image (use edit_image_mesh) or re-roll it (use regenerate_image_mesh). Effects: Creates a new project in the user's Flow Studio library. Spends credits from the user's Flow Studio license. Returns: generationId, the input to edit_image_mesh, regenerate_image_mesh and (as sourceGenerationId) create_3d_model_from_image_mesh. Shows: the image viewer with progress and the finished images.

delete_generation_mesh

ChatGPT
Permanently deletes one mesh generation. Use when: the user explicitly asks to delete a mesh generation. Do not use when: the user wants a different take on a result (use regenerate_image_mesh) or to change it (use edit_image_mesh). Effects: Removes the generation from the user's Flow Studio library. The deletion is permanent and can't be undone. Returns: generationId and deleted.

edit_image_mesh

ChatGPT
Edits one variant of an existing generated image using a text instruction (e.g. 'make the sky red'). Use when: the user wants a specific change to a Flow Studio image, including a follow-up like 'make it red' after an image from this server. Do not use when: a new variant with no new instruction (use regenerate_image_mesh); or a shape change asked of a 3D draft or model, before an explicit yes to a new image and 3D model. Effects: Spends credits from the user's Flow Studio license. Returns: generationId of the edit, the input to edit_image_mesh, regenerate_image_mesh and (as sourceGenerationId) create_3d_model_from_image_mesh. Shows: the image viewer with progress and the finished images.

get_action_ai_model_options_mesh

ChatGPT
Returns the AI models the user can pick for one mesh action, with cost, time estimate, limits, the default and the user's saved preference. Use when: the user asks which AI models they can use for a mesh action, what a model costs or how long it takes, or a model choice for a mesh tool needs checking. Do not use when: the user asks for the status or outputs of a generation (use get_generation_mesh), or what they can do next with a finished generation (use get_generation_next_step_options_mesh). Returns: models with modelKey (the aiModel input of the action's tool), displayName (Flow Studio UI name), cost, timeEstimationSeconds, tiersRestricted and enumConstraints; texturingModels when the action takes one; defaultModelKey, userPreferenceModelKey and suggestedParameters.

get_generation_mesh

ChatGPT
Returns the current state of one mesh generation, with progress and output URLs when complete. Use when: the user asks for the status or progress of a generation, or a follow-up step needs its task ids or output URLs. Do not use when: the user wants a new image or 3D model (use create_image_from_text_mesh or create_3d_model_from_text_mesh), to see one existing generation (use show_generation_mesh), or what they can do next with a finished generation (use get_generation_next_step_options_mesh). Returns: state, progressPercent, isTerminal, tasks[] (taskId, the input to remesh_and_retexture_draft_mesh, remesh_finished_model_mesh and get_generation_download_link_mesh), outputImageId and signed output URLs when complete. Data only: this tool has no viewer.

get_generation_next_step_options_mesh

ChatGPT
Returns the next-step options for a mesh generation: the choices, their defaults and what each one does. Use when: a mesh generation has completed, or the user asks what they can do next with an image or 3D model. Do not use when: the user asks for the status or outputs of a generation (use get_generation_mesh), or which AI models an action can use (use get_action_ai_model_options_mesh). Returns: userOptionsMarkdown (the choices written for the user), scenario, recommendedTool (the tool for the main path), criteria, otherActions and suggestedDefaults.

get_user_ai_model_preferences_mesh

ChatGPT
Returns the user's saved default AI model per generation type (the same defaults as the Flow Studio web app). Use when: the user asks which AI models are their saved defaults. Do not use when: the user wants to change a default (use set_user_ai_model_preference_mesh) or pick a model for one action (use get_action_ai_model_options_mesh). Returns: the saved default model per generation type.

list_licenses_billing

ChatGPT
Lists the user's Flow Studio licenses with licenseId, plan name and remaining/total credits. Use when: the user asks about their plans or credits. Do not use when: the user wants to start a generation or know whether their plan allows it (use check_generation_access_billing). Returns: one entry per license with licenseId (the licenseId input to the generation tools), plan name, remaining and total credits, and hasCredits.

list_project_generations_mesh

ChatGPT
Lists the generations inside one mesh project, with preview image URLs on tasks when available. Use when: the user wants every generation of one project, beyond the recent ones list_projects_mesh shows, or pages through them. Do not use when: the user wants their library, a list of projects, or one project by name (use list_projects_mesh with search; its viewer shows that project's recent generations). Returns: items with ids, label (e.g. "Image 3", the name on the tile), type and tasks, plus paging fields. Shows: a thumbnail gallery with the project name and each generation's preview, tiles labelled ("Image 3"). The viewer's model context holds the generationId and taskId of the tiles on screen and the open tile.

list_projects_mesh

ChatGPT
Lists mesh library projects with optional recent generation previews embedded (same as Flow Studio My Library). Use when: the user asks for their Flow Studio library or projects, for one project by name (search; the viewer shows its recent generations), or for projects by asset type. Do not use when: the user wants every generation of one project (use list_project_generations_mesh) or one generation opened (use show_generation_mesh). Returns: items with project id, name, generation count and recent generations, plus paging fields. Shows: project cards with previews and a "Show more" button. Tiles are labelled per project row ("Image 2"). The viewer's model context holds the generationId and taskId of the tiles on screen and the open tile.

regenerate_image_mesh

ChatGPT
Regenerates one variant of an existing generated image with the same prompt and no new instruction (a re-roll). Use when: the user wants a different take on the same prompt of a Flow Studio image (e.g. 'try again', 'another one'). Do not use when: the user wants a specific change (use edit_image_mesh). Effects: Spends credits from the user's Flow Studio license. Returns: generationId of the new image, the input to edit_image_mesh, regenerate_image_mesh and (as sourceGenerationId) create_3d_model_from_image_mesh. Shows: the image viewer with progress and the finished images.

remesh_and_retexture_draft_mesh

ChatGPT
Remeshes a completed draft 3D generation with Max, Fixed or Adaptive polycount, and optionally textures it. Use when: the user wants to refine a completed draft 3D model: a final polycount and topology, with or without texture. Do not use when: the mesh is finished, not a draft (use remesh_finished_model_mesh or retexture_finished_model_mesh); or the user adds or changes parts (clothing, props): no tool changes the shape; that takes a new image and create_3d_model_from_image_mesh, both spending credits. Effects: Spends credits from the user's Flow Studio license. Creates a new generation. Returns: generationId, the input to get_generation_mesh and get_generation_next_step_options_mesh; projectId, state and the licenseId billed. Shows: the 3D viewer with progress and the finished model.

remesh_finished_model_mesh

ChatGPT
Remeshes a finished (non-draft) mesh to a new polycount and topology. Use when: the user wants a different polycount or topology on a finished mesh. Do not use when: the mesh is a draft, or the user wants Max polycount (use remesh_and_retexture_draft_mesh); or the user adds parts (clothing, props): no tool changes the shape; that takes a new image and create_3d_model_from_image_mesh, both spending credits. Effects: Spends credits from the user's Flow Studio license. Creates a new generation. Returns: generationId, the input to get_generation_mesh and get_generation_next_step_options_mesh; projectId, state and the licenseId billed. Shows: the 3D viewer with progress and the finished model.

retexture_finished_model_mesh

ChatGPT
Retextures a finished (non-draft) mesh from a text prompt or a reference image. Use when: the user wants a new texture on a finished mesh. Do not use when: the mesh is a draft (use remesh_and_retexture_draft_mesh); or the user adds or changes parts (clothing, props): no tool changes the shape; that takes a new image and create_3d_model_from_image_mesh, both spending credits. Effects: Spends credits from the user's Flow Studio license. Creates a new generation. Returns: generationId, the input to get_generation_mesh and get_generation_next_step_options_mesh; projectId, state and the licenseId billed. Shows: the 3D viewer with progress and the finished model.

search_generations_mesh

ChatGPT
Finds images and 3D models the user made, by project name or type, to open one with show_generation_mesh. Shows nothing to the user. Use when: the user wants to open one image or 3D model they made, named by its project or described ("the last dog image"). Do not use when: the user wants to browse projects or generations, also by name (use list_projects_mesh with search), or a known generation's status (use get_generation_mesh). Returns: matching projects with their newest generations, newest first; generationId is the input to show_generation_mesh.

set_user_ai_model_preference_mesh

ChatGPT
Sets the user's saved default AI model for one mesh action, the same setting as the model dropdown in the Flow Studio web app. Use when: the user wants to change their default AI model for a mesh action. Do not use when: the user wants a different model for one generation only (the aiModel parameter of the create tool covers that). Effects: Replaces the saved default for that action, in chat and in the Flow Studio web app; later generations of that action use it. Spends no credits. Returns: the updated map of saved default models per generation type.

show_generation_mesh

ChatGPT
Opens one existing generation, an image or a 3D model, in a single-result viewer. Use when: the user wants to see one existing generation whose id is known. Do not use when: the user wants to find a generation (use search_generations_mesh), browse the library (use list_projects_mesh), only check progress (use get_generation_mesh) or just created it (the create tool's viewer shows it). Returns: generationId, kind, type, state and projectId. Shows: the image or the 3D model of that generation, with progress while it is running.

Capabilities

WritesInteractive

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/autodesk-flow-studio"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/autodesk-flow-studio)