Godot MCP Runtime vs Ludo AI Game Assets
A side-by-side look at the Godot MCP Runtime and Ludo AI Game Assets MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose Godot MCP Runtime for gaming work as a local process; choose Ludo AI Game Assets for gaming over a hosted endpoint. You can also run both side by side in one mcpServers block.
Which MCP server should you choose?
Choose Godot MCP Runtime when
- You need dedicated capabilities in the Gaming domain.
- You prefer a local stdio subprocess transport.
- Its auth model fits you: No auth required.
- You want tools like
launch_editor, run_project, attach_project.
Choose Ludo AI Game Assets when
- You need dedicated capabilities in the Gaming domain.
- You prefer a remote HTTP/SSE transport.
- Its auth model fits you: API key required.
- You want tools like
validateApiKeyEndpoint, createImage, generateWithStyle.
Feature & specification comparison
Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | Gaming | Gaming |
| Quality signal | 69/100 · Great | 61/100 · Good |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Remote HTTP/SSE |
| Auth | No auth required | API key required |
| Pricing | Free | Not declared |
| Required env vars | None listed | None listed |
| Verified owner | Community listing | Community listing |
| GitHub stars | 95 | 7 |
| Directory activity | 8 views 0 installs 0 upvotes | 1 views 0 installs 0 upvotes |
| Tools | 39 | 44 |
| Full listing | View Godot MCP Runtime | View Ludo AI Game Assets |
Tools & capabilities
Godot MCP Runtime tools 39
launch_editorOpen the Godot editor GUI for a project for the human user. Use only when the user explicitly asks to "open the editor"; for any agent-driven work, use the headless scene/node tools (add_node, set_node_properties, etc.) instead - the editor cannot be controlled programmatically. Returns plain-text confirmation after spawning the editor process. Errors if projectPath has no project.godot.run_projectSpawn a Godot project as a child process with stdout/stderr captured. Required before take_screenshot, simulate_input, get_ui_elements, run_script, or get_debug_output. Set profiling: true at launch to enable the profiler tools. Use attach_project for one you launched yourself. Verifies MCP bridge readiness before returning success. Returns status with the assigned bridge port. Call stop_project when done. Errors if projectPath is not a Godot project or another session is already active.attach_projectInject the MCP bridge into a Godot process you launch yourself, then wait up to 20s for the bridge to start listening and up to 45s total once it has, so a large project's cold start is absorbed; a port that listens but answers no ping gives up sooner. Call BEFORE Godot launches - Godot reads autoloads only at process start, so a late call returns "bridge did not respond." Recommended pattern: kick off the Godot launch in parallel with this call so the wait absorbs startup. Prefer run_project unless MCP must not spawn Godot. Only one attach session is supported per project at a time (attach mode has no env-var channel, so port and token must be baked into the one shared script); a second attach_project on a project another session already attached to is refused, naming that session. Returns plain-text status with the resolved bridge port. Call detach_project or stop_project when done.detach_projectClear attached-mode runtime state and remove the injected McpBridge autoload. Does NOT stop the manually launched Godot process - that stays running. Use after attach_project when you are done driving the game from MCP. For spawned sessions (run_project), use stop_project instead. Mostly optional now: when the bridge disconnects (you closed Godot), the next runtime tool call probes once and ends the attached session itself, removing the autoload. Calling it afterwards still succeeds idempotently, wording the message to distinguish "an attached session existed and already ended" from "this server never attached to a project". Returns: message confirming detach plus externalProcessPreserved (always true here - that is the point of detach vs stop_project). Errors only when a spawned session is what is active; use stop_project for those.get_debug_outputGet captured stdout/stderr from a spawned Godot project. Use whenever runtime tools fail unexpectedly - script errors, missing nodes, and crash backtraces all surface here. Still works after the process exits or crashes: the session clears itself on exit but the captured logs are retained until stop_project. Requires run_project (not attach_project; attached mode does not capture output). Returns: output/errors (last `limit` lines each, default 200), running (false after exit, null when attached), exitCode after exit, attached:true with empty arrays in attached mode.stop_projectStop the spawned Godot project and clean up bridge state. Call when done with runtime testing, even after a crash, and even if you closed the Godot window yourself: it frees the process slot and clears the flag blocking scene-editing tools. A process that exited on its own already removed the bridge autoload at that moment, and this still succeeds - it reports alreadyExited:true with the exit code and the logs captured before the exit, and leaves a finished profiler capture readable. Attached sessions detach without killing the external process. Returns: message, mode, externalProcessPreserved, alreadyExited, exitCode (already-exited case), and condensed finalOutput/finalErrors (capped at 200); get_debug_output has the full log. Errors only when there is no session and no exited process to report.take_screenshotCapture a PNG of the running viewport. responseMode: preview (default - saves full PNG, returns bounded inline preview at 960x540), full (full inline PNG; use for small text or pixel-level inspection), path_only (saved-path only, no inline image). Saved under .mcp/godot-runtime/screenshots/ (persists after stop_project). Returns: inline image block (full/preview modes), plus path and size of the saved PNG; previewPath/previewSize in preview mode; warnings for non-fatal runtime errors. Errors if no session or bridge times out (default 10000ms).simulate_inputSimulate sequential input in a running project and report what each action did. Action `type`: key, mouse_button, mouse_motion, click_element, action, text, wait. For key/mouse_button/action, omit `pressed` to tap (press+release); set it to hold or release. click_element resolves by node path/name (see get_ui_elements), not visible text. Returns: results[] per action with ok, timing, signals fired, the Control hit, UI `changes` (appeared/disappeared/changed), `watch` samples, and `errors` from input handlers (spawned sessions only). Invalid batches inject nothing; a runtime failure stops the batch and skips the rest.get_ui_elementsWalk the running scene tree and return all Control nodes with positions, sizes, types, and text content. Always call this before simulate_input click_element actions to discover valid element names and paths. Requires an active runtime session (run_project or attach_project). visibleOnly defaults true; pass false to include hidden Controls. filter narrows by class. Returns: elements[] with path/type/rect/visible plus optional text/disabled/tooltip.run_scriptExecute a custom GDScript in the live running project with full scene tree access. Requires an active runtime session. Script must extend RefCounted and define func execute(scene_tree: SceneTree) -> Variant. Return values are JSON-serialized (primitives, Vector2/3, Color, Dictionary, Array, and Node path strings). Use print() for debug output - it appears in get_debug_output, not in the result. In spawned mode, stderr runtime errors escalate to errors (when the script returns null) or surface as warnings. Returns: { success, result, warnings?, tip? } where result is the JSON-serialized return value of execute().list_autoloadsList all registered autoloads in a project with paths and singleton status. Use first when diagnosing headless failures - broken autoloads crash all headless ops, so this tells you what is loaded. No Godot process required (reads project.godot directly). Returns: [{ name, path, singleton }].add_autoloadRegister a new autoload in a project. autoloadPath accepts "res://..." or a project-relative path (auto-prefixed). singleton defaults true (accessible globally by name). No Godot process required. Warning: autoloads initialize in headless mode - a broken script will crash every subsequent headless op; validate before adding. Returns plain-text confirmation with the registered name, path, and singleton flag. Errors if an autoload with the same name already exists; use update_autoload to modify.- +27 more on the listing page
Ludo AI Game Assets tools 44
validateApiKeyEndpointValidates an API key. Returns 200 if valid, 403 if invalid.createImageGenerate game-art images from a text prompt alone, selecting an image_type (e.g. sprite) and optionally art_style, perspective, and aspect_ratio. The job result is an array of image results, each with a url; request n (1-8) to control how many variations come back. Because it generates purely from text it takes no source image, so there is no upload size limit to trip. Credits are held when the job is accepted and refunded if it fails or is cancelled; the charge scales with the number of images produced. Use createImage to make new images from scratch; use generateWithStyle to match a reference image's art style, editImage to modify an existing image, and removeBackground to cut out a subject. Pass an optional request_id to tag the results so you can retrieve them later via listGenerations (type image). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: This endpoint consumes 0.5 credits per result.generateWithStyleGenerate new images that match the visual style of a reference image: supply a style_image (URL or base64) plus a text prompt describing what to create and an optional image_type (defaults to sprite). The job result is an array of image results, each with a url; request n (1-4) to control the number of variations. The style_image is uploaded and validated, and an image larger than 15MB is rejected with HTTP 400. Credits are held when the job is accepted and refunded if it fails or is cancelled; the charge scales with the number of images produced. Use this instead of createImage when style consistency with an existing asset matters; use editImage to alter the content of a specific image rather than borrow its style, and removeBackground to isolate a subject. Pass an optional request_id to tag the results so you can retrieve them later via listGenerations (type image). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: This endpoint consumes 0.5 credits per result.editImageModify an existing image according to text instructions: supply a source image (URL or base64) and a prompt describing the changes (e.g. "add clouds", "warmer color scheme"), with an optional reference_image for extra style or content guidance. To change only one area, pass a mask (white = change, black = keep, the same size as the image): mask_mode "replace" (default) redraws that area from the prompt, "erase" removes what is there. Without a mask the whole image is re-rendered. Never ask for a transparent background in the prompt: it often comes back as a grey-and-white checkerboard painted into the pixels, which removeBackground cannot remove. A sprite keeps its transparency without being asked (see image_type). The job result is an array of image results, each with a url; request n (1-4) to control the number of edited variations. Provided images are uploaded and validated, and any image larger than 15MB is rejected with HTTP 400. Credits are held when the job is accepted and refunded if it fails or is cancelled; the charge scales with the number of images produced. Use editImage to transform a specific existing image; use createImage to generate from text alone, generateWithStyle to borrow a reference's art style, and removeBackground for the dedicated background-removal case. Pass an optional request_id to tag the results so you can retrieve them later via listGenerations (type image). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: This endpoint consumes 0.5 credits per result; 1 credit per result with a mask.animateSpriteAnimate a static sprite into a spritesheet driven by a motion text prompt (image-to-spritesheet): supply an initial_image (URL or base64) plus a motion_prompt like "walking" or "attack slash", and optionally a final_image to interpolate between a start and end frame. Describe the motion exactly and unambiguously, but do not over-describe it: an action the model already knows is one phrase, not a sequence of steps, and the character, art style, background and camera come from the image, not the prompt. For animations driven by up to three keyframes (including a middle frame), use animateSpriteKeyframes instead. The job result is a single sprite result: `spritesheet_url` (the sheet image), `video_url` (an mp4 of the animation - pass it as `video` to transferMotion or as `spritesheet_video_url` to createSpriteAudio; editSpritesheet takes `spritesheet_url`), `num_frames`/`num_cols`/`num_rows` (the grid layout), and, when requested, `gif_url`, `individual_frame_urls` and `spritesheet_with_background_url`. The hydra and hydra-plus models also return `audio_url`, a sound effect for the animation - there is no need to call createSpriteAudio afterwards; forge and forge-pixel produce no audio. The chosen model must support sprite animation and the duration must be valid for it; incompatible model/duration combinations return HTTP 400. Credits are held for the duration you requested when the job is accepted; the final charge is rate × seconds of video generated (legacy models keep their minimum charge), never more than for the duration you requested, and the difference (or everything, if the job fails or is cancelled) is refunded. The returned sheet is then trimmed to its best loop and can run slightly shorter than the video; that trim is not deducted. Choosing a tool: animateSprite (this one) when you can describe the motion in text; transferMotion when you want an exact motion copied from a reference video or one of the named presets from listAnimationPresets (e.g. a standard walk or attack cycle); generatePose first when the source image is not yet in the pose the animation should start from. Omit `model` to run on hydra (the default) - see the `model` field for the per-model rates. Pass an optional request_id to tag the result so you can locate it later via listGenerations (type spritesheet). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: credits/s × seconds, per model: Hydra 3/s (shortest duration 3s, so 9 credits minimum), Hydra+ 4/s (shortest duration 3s, so 12 credits minimum), Forge 1.5/s, Forge Pixel 1.5/s, Blitz 1.9/s (min charge 4 credits) [LEGACY], Eagle 2.6/s (min charge 4 credits) [LEGACY], Eagle with Audio 3.1/s (min charge 4 credits) [LEGACY]; [LEGACY] models are scheduled for removal - do not use them for new work; see this endpoint's full pricing table in the API docs.animateSpriteKeyframesAnimate a sprite through up to three fixed keyframes - initial_image, middle_image and final_image (each a URL or base64) - producing a spritesheet that interpolates through the provided frames in order. At least one of initial_image or middle_image is required (a final_image alone has nothing to anchor the animation); any keyframe may be omitted. The motion_prompt is optional here - when omitted, the motion is derived purely from the keyframes. Runs on hydra (default; also returns `audio_url`), hydra-plus (hydra with finer detail, also returns `audio_url`), forge, or forge-pixel for pixel-art sprites - the models that support a middle keyframe; any other model returns HTTP 400. For just a start and end frame, animateSprite with `final_image` does the same job. The job result is the same sprite result as animateSprite (`spritesheet_url`, `video_url`, grid fields, optional GIF and frame URLs, `audio_url` on hydra and hydra-plus). Credits are held for the duration you requested when the job is accepted; the final charge is rate × seconds of video generated, never more than for the duration you requested, and the difference (or everything, if the job fails or is cancelled) is refunded. The returned sheet is then trimmed to its best loop and can run slightly shorter than the video; that trim is not deducted. Use animateSprite instead for the classic single-image + text-prompt animation. Pass an optional request_id to tag the result so you can locate it later via listGenerations (type spritesheet). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: credits/s × seconds, per model: Hydra 3/s (shortest duration 3s, so 9 credits minimum), Hydra+ 4/s (shortest duration 3s, so 12 credits minimum), Forge 1.5/s, Forge Pixel 1.5/s; see this endpoint's full pricing table in the API docs.generatePoseRe-pose an existing sprite into a new target pose while preserving the character, taking a source image (URL or base64), a pose name (or "Other" with a free-text description), and an optional n (1-4) for how many variations to produce. Only works with sprite image types (not icons, screenshots, etc.). The job result is an array of pose results, each containing the generated image url, the pose and description used, and a suggested motion_prompt tuned for that pose. Credits are held when the job is accepted and refunded if it fails or is cancelled; the charge scales with the number of images generated. This is typically the first step before animating: call generatePose to set the character's pose, then feed the result (and its suggested motion_prompt) into animateSprite for the best animation quality; use rotateSprite instead when you want to change the camera angle rather than the pose. Pass an optional request_id to tag the results so you can locate them later via listGenerations (type spritesheet). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: This endpoint consumes 0.5 credits per result.rotateSpriteRe-render an existing sprite from a different camera viewpoint while keeping the same character and pose, taking a source image (URL or base64), a required camera_rotation azimuth (one of 0, 45, 90, 135, 180, -135, -90, -45 degrees), an optional camera_elevation (0, 30, or 60 degrees; omit to keep the current elevation), and an optional n (1-4) for the number of variations. Only works with sprite image types (not icons, screenshots, etc.). The job result is an array of rotate-sprite results, each with the generated image url and the camera_rotation and camera_elevation that were applied. Credits are held when the job is accepted and refunded if it fails or is cancelled; the charge scales with the number of images generated. Use this to produce alternate view angles of a character; use generatePose instead to change the character's pose rather than the camera, and animateSprite or transferMotion to bring a sprite to life. Pass an optional request_id to tag the results so you can retrieve them later via listGenerations (type spritesheet). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: This endpoint consumes 0.5 credits per result.transferMotionTransfer motion from a reference video or a named animation preset onto a static sprite image, producing an animated spritesheet that mimics the reference movement. Provide the sprite as image (URL or base64) plus either a video URL or a preset_id together with perspective and direction (all three from listAnimationPresets; if both video and preset_id are sent the video wins). The job result is the same sprite result as animateSprite: `spritesheet_url`, `video_url`, grid fields, optional GIF / frame / with-background URLs, and `audio_url` when the model is hydra. `duration` defaults to 1.5s where the chosen model offers it, otherwise to that model's shortest (hydra, the default, starts at 3s) - and a longer reference clip or preset is compressed to fit, so pass the preset's own `duration` (returned by listAnimationPresets) to keep its timing. It returns HTTP 400 if neither a video nor a complete preset_id/perspective/direction triple is supplied, if the named preset, perspective, or direction cannot be resolved, or if the model/duration combination is invalid. Credits are held for the duration you requested when the job is accepted; the final charge is rate × seconds of video generated (which follows the reference clip or preset; legacy models keep their minimum charge), never more than for the duration you requested, and the difference (or everything, if the job fails or is cancelled) is refunded. The returned sheet is then trimmed to its best loop and can run slightly shorter than the video; that trim is not deducted. Use this when you have an existing motion clip or preset to copy; prefer animateSprite to generate animation purely from a text prompt. Omit `model` to run on hydra (the default - most capable, and returns audio); pick forge for a cheaper run on simple motion, presets and matching poses; the `model` field lists rates. Pass an optional request_id to tag the result so you can retrieve it later via listGenerations (type spritesheet). Async generation job: returns `{id, status}` - poll `getApiJob` (job and credit contract: see the server instructions). Credits: credits/s × seconds, per model: Hydra 3/s, Forge 2/s, Forge Pixel 2/s, Tango 4/s (min charge 4 credits) [LEGACY]; [LEGACY] models are scheduled for removal - do not use them for new work; see this endpoint's full pricing table in the API docs.getDocsBrowse or read Ludo's own documentation in full - how to choose a sprite animation mode, when to use Generate Before / Generate After, how margins behave, which model suits a job, and each generator's known limitations. To answer a specific question, call searchDocs first: it returns just the sections that match. Use getDocs to see what documentation exists, or to read a whole document or named sections. This is the same documentation the Ludo web app shows its users, so it occasionally describes buttons rather than parameters; the substance applies to the API and MCP surfaces just the same. To browse, call it with NO parameters to get a table of contents - every document with its id, label and section titles, and no bodies - then call it again with `doc` (and optionally `sections`) to read only what you need. Fetching a whole document can return tens of thousands of characters, so prefer naming the sections, using titles copied from the table of contents. Section titles match ignoring case, spacing and punctuation: when only some requested titles exist you receive those sections plus `unmatched_sections` and `available_sections`, and when none exist (or `doc` is unknown) the call returns 400 listing the valid values so you can retry once. This is a free discovery endpoint: it does not charge credits and does not queue a job.searchDocsSearch Ludo's own documentation with a plain-language question and get back only the few sections that answer it - the fastest way to learn how a feature is meant to be used before you generate with it (how to pick a sprite animation mode or model, how margins behave, what something costs, known limitations). Start here rather than reading whole documents. Each result carries `doc` and `section`, which you can pass straight to getDocs to re-read that section, and the section's full markdown `content`. Results are best first; weak matches are left out, so an empty `results` list means the documentation does not cover the question - rephrase it, or call getDocs with no parameters to browse the table of contents. This is the same documentation the Ludo web app shows its users, so it occasionally describes buttons rather than parameters; the substance applies to the API and MCP surfaces just the same. Returns up to `n` sections (default 3, max 10). If it answers 503 the search backend is briefly unavailable: call getDocs instead rather than retrying in a loop. This is a free discovery endpoint: it does not charge credits and does not queue a job.listAnimationPresetsList the available animation presets along with their perspectives and the eight supported compass directions (N, NE, E, SE, S, SW, W, NW). Synchronous GET with no request body: it returns an animations array (each with id, name, category, description, duration, and preview_url), a deduplicated perspectives array, and the directions list. This is a free discovery endpoint and does not charge credits. Use it to obtain the preset_id, perspective, and direction values that transferMotion needs, and to find motion preset names you can reference when animating; pair it with transferMotion (to apply a preset onto a sprite), animateSprite (text-prompt animation), or animate3DModelPreset (apply any preset by id to a rigged 3D model).- +32 more on the listing page
Ready-to-paste client configs
Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
Godot MCP Runtime
{
"mcpServers": {
"erodenn-godot-mcp-runtime": {
"command": "npx",
"args": [
"-y",
"godot-mcp-runtime"
]
}
}
}
Ludo AI Game Assets
{
"mcpServers": {
"ludo-ai-game-assets": {
"url": "https://mcp.ludo.ai/mcp"
}
}
}
Frequently asked questions
Godot MCP Runtime is categorized under Gaming and uses a local stdio subprocess. Ludo AI Game Assets belongs to Gaming and uses a remote HTTP/SSE transport. Pick Godot MCP Runtime for gaming work and Ludo AI Game Assets for gaming.
Copy the mcpServers block from the config section on this page into your client's config (for example claude_desktop_config.json or ~/.cursor/mcp.json), then fully restart the client.
Godot MCP Runtime is listed as free. Auth: no auth required. Ludo AI Game Assets has no declared pricing — check the project's own site. Auth: api key required. If a server needs environment variables such as API keys, define them under the env key in your client's MCP config.
Yes. MCP clients support multiple servers: add erodenn-godot-mcp-runtime and ludo-ai-game-assets as separate keys in the same mcpServers object, and your assistant will call tools from either one.
On AllMCPs, Godot MCP Runtime has a quality signal of 69/100 (Great) with 8 views, 0 installs and 95 GitHub stars. Ludo AI Game Assets has 61/100 (Good) with 1 views, 0 installs and 7 GitHub stars.