R3
REMOTE · MCP.R3JS.ORG
Build 3D web apps: physics, audio, music, Blender import. Live edits, undo, versions, publishing.
Available components
Last scanned (7 hours ago)
Scores higher than 99% of remote servers
Reachable. Last reached 21 Nov 2026.
Category Developer Tools
How this component scores in each security and reliability category. Every signal is checked automatically against the live server, and we only credit what we can confirm. How we score → Why this is hard to score →
Endpoint Security97
- The endpoint's TLS certificate is valid, in date, and uses a strong key. View diagnostics → Pass
- Authorisation is enforced on tool calls, advertised via RFC 9728 protected-resource metadata. Discovery is public, which costs nothing: no tool can be invoked without a token. View diagnostics → Pass
- HTTPS is enforced; there's no plaintext access path. View diagnostics → Pass
- The HSTS (Strict-Transport-Security) header is present. View diagnostics → Pass
- DNSSEC check failed: this domain isn't protected by DNSSEC. See how to fix → View diagnostics → Fail
- The authorisation server supports Client ID Metadata Documents, the current MCP client-registration mechanism. View diagnostics → Pass
Transport & Reachability100
- Verified streamable-http transport via a live MCP handshake. View diagnostics → Pass
Schema Quality & AI Usability77
- 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).Pass
- AI-judged instruction clarity (excellent).Pass
- Context-footprint check failed: tool/resource definitions use about 27868 tokens (~238/item across 117 items; 80 tools + 37 resources), over budget; trim descriptions and params. See how to fix → Fail
- Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management98
- Stability check failed: schema churn in the 30 days we've observed: 1 tool removals, 0 breaking changes, 0 auth/transport breaks, 19 additions. See how to fix → Fail
Tool Coverage100
- 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
- 100% of tool parameters carry a description.Pass
- Structured output schemas are declared (98% of tools); any adoption earns full credit.Pass
Tool Safety97
- No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
- 7 of 8 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "r3_publish_settings_read" implies "publish" and declares readOnlyHint instead, contradicting what its own name says it does. See how to fix → Partial
- An AI judge read all 82 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
- Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.Pass
- Supports UI / widget rendering.Pass
Checks we ran (14)
- TLS certificate
- Authorization
- HTTPS
- HSTS header
- DNSSEC
- Transport
- Schema quality
- Stability
- Tool coverage
- Injection markers
- Destructive annotations
- Judged manipulation
- MCP protocol
- Capabilities
Does not apply to this hosted endpoint (2)
- Supply Chain SecurityScored on package channels only. This is a hosted endpoint, so there is nothing to install or inspect.
- Provenance & TransparencyScored on package channels only. This is a hosted endpoint, so there is nothing to install or inspect.
The score does not measure everything. What the score does not measure →
Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.
- 21 Nov 26 0
- Server version: 3.222.0 → 3.223.0 functional
- Server version: 3.221.0 → 3.222.0 functional
- Server version: 3.220.0 → 3.221.0 functional
- 20 Nov 26 0
- Server version: 3.219.0 → 3.220.0 functional
- Server version: 3.218.0 → 3.219.0 functional
- 19 Nov 26 0
- New resource “HOWTO_TEXT.md” functional
- Server version: 3.217.0 → 3.218.0 functional
- Server version: 3.216.0 → 3.217.0 functional
- Server version: 3.215.0 → 3.216.0 functional
- Server version: 3.214.0 → 3.215.0 functional
- Server version: 3.213.0 → 3.214.0 functional
- Server version: 3.212.0 → 3.213.0 functional
- Server version: 3.210.0 → 3.212.0 functional
- Server version: 3.208.0 → 3.210.0 functional
- 18 Nov 26 0
- Server version: 3.207.0 → 3.208.0 functional
- Server version: 3.206.0 → 3.207.0 functional
- Server version: 3.205.0 → 3.206.0 functional
- Server version: 3.204.0 → 3.205.0 functional
- Server version: 3.202.0 → 3.204.0 functional
- Server version: 3.200.0 → 3.202.0 functional
- Server version: 3.199.1 → 3.200.0 functional
- 17 Nov 26 0
- Server version: 3.198.1 → 3.199.1 functional
- Server version: 3.197.0 → 3.198.1 functional
- Server version: 3.196.0 → 3.197.0 functional
- 16 Nov 26 0
- Resource “HOWTO_GAMES.md” was removed ▼ functional
- Server version: 3.195.0 → 3.196.0 functional
- Server version: 3.194.0 → 3.195.0 functional
- 15 Nov 26 0
- Server version: 3.193.0 → 3.194.0 functional
- Server version: 3.192.0 → 3.193.0 functional
- Server version: 3.191.0 → 3.192.0 functional
- 14 Nov 26 0
- New resource “HOWTO_GAMES.md” functional
- Server version: 3.190.0 → 3.191.0 functional
- Server version: 3.188.0 → 3.190.0 functional
- Server version: 3.186.0 → 3.188.0 functional
Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.
Captured 21 Nov 2026 · Probed https://mcp.r3js.org/mcp
TLS valid
Negotiated TLS 1.3 with TLS_AES_128_GCM_SHA256 .
| Subject | Issuer | Valid from | Valid until | Key | Signature | Serial |
|---|---|---|---|---|---|---|
| CN=mcp.r3js.org | CN=YE2,O=Let's Encrypt,C=US | 26 Oct 2026 | 24 Jan 2027 | ECDSA 256 | ECDSA-SHA384 | 67f99411823edf75d7cbcba072a4f7d33c6 |
| SANs: mcp.r3js.org | ||||||
| CN=YE2,O=Let's Encrypt,C=US (CA) | CN=Root YE,O=ISRG,C=US | 3 Sept 2025 | 2 Sept 2028 | ECDSA 384 | ECDSA-SHA384 | 4df3b15dd6c0784c507cd37b58e6f115 |
| CN=Root YE,O=ISRG,C=US (CA) | CN=ISRG Root X2,O=Internet Security Research Group,C=US | 13 May 2026 | 2 Sept 2032 | ECDSA 384 | ECDSA-SHA384 | 872165fc34b6e5fba8add5b3705fb53a |
| CN=ISRG Root X2,O=Internet Security Research Group,C=US (CA) | CN=ISRG Root X1,O=Internet Security Research Group,C=US | 13 May 2026 | 2 Sept 2032 | ECDSA 384 | SHA256-RSA | 6c8f1dc727c7117f7baf853ac980f9cd |
Background: What to check on a remote MCP endpoint →
DNSSEC insecure
Validation of mcp.r3js.org. — Not signed
| Zone | DS | Keys | Algorithms | Outcome |
|---|---|---|---|---|
| . | Trust anchor | 20326, 38696 | 8, 8 | Verified |
| org. | DS published | 725 | 8 | Verified |
| r3js.org. | No DS | Unsigned (proven) parent-signed NSEC/NSEC3 proves an unsigned delegation |
Authentication Enforced and verified
The endpoint asked for a token and published valid RFC 9728 metadata describing how to get one.
| Result | Enforced and verified |
|---|---|
| Enforced | On tool calls |
| HTTP status | 200 |
WWW-Authenticate challenge Bearer realm="r3", resource_metadata="https://mcp.r3js.org/.well-known/oauth-protected-resource/mcp", scope="project:read project:write code:write editor:dispose editor:run project:delete:created asset:read asset:write publish:preview support:write roadmap:suggest roadmap:vote"
Bearer realm="r3", resource_metadata="https://mcp.r3js.org/.well-known/oauth-protected-resource/mcp", scope="project:read project:write code:write editor:dispose editor:run project:delete:created asset:read asset:write publish:preview support:write roadmap:suggest roadmap:vote" | Header | Value |
|---|---|
| strict-transport-security | max-age=63072000; includeSubDomains; preload |
| x-content-type-options | nosniff |
| x-frame-options | DENY |
Protected resource metadata
| Document | https://mcp.r3js.org/.well-known/oauth-protected-resource/mcp |
|---|---|
| Retrieved | Yes |
| Resource | https://mcp.r3js.org/mcp |
| Authorisation server | https://mcp.r3js.org |
Background: How OAuth 2.1 works in the 2026 MCP spec →
Transports 2 probes
| Transport | URL | Outcome | Status | Location |
|---|---|---|---|---|
| streamable-http | https://mcp.r3js.org/mcp | Verified | 200 | |
| http (plaintext) | http://mcp.r3js.org/mcp | HTTPS enforced | 301 | https://mcp.r3js.org/mcp |
The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →
r3_asset_delete Delete an image, model, audio file or Blender model ~477
Delete one of the account's images, models, audio files, Blender models (kind blender: a whole uploaded .blend with every version; its meshes in use stay) or texture sets (kind textureSet: the set and its maps, r3_texture_create made it) — the same reference-guarded delete as the R3 app. With asset:delete:created, only an asset YOU created (this agent key or app connection; today: images made with r3_image_create); with asset:delete:owned, any of the user's assets, and only when the user asked for it: pass owned: true for one you did not create (an app connection is then asked for that right). Two calls, every time: without `confirm` nothing is deleted, and the answer says what would go and which of the user's projects use it; the same call with `confirm: true`, when that is what the user wants, deletes it. An asset another user's project uses can never be deleted. A model takes the images only it used with it, an HDR / EXR image its preview image. There is no trash and no undo. Ids come from r3_image_list, r3_audio_list and r3_blender_list.
| Name | Type | Req | Description |
|---|---|---|---|
| confirm | boolean | – | Leave out on the first call (it only reports). true deletes — after the user has seen the report. |
| deleteReleases | boolean | – | With confirm: also move the user's non-live, UNPINNED releases that use it to the 7-day trash (needs release:delete; only when the user asked). Without it those releases are kept: for an image, model… |
| id | string | yes | The asset's id, as the list tools report it. |
| kind | string | yes | What the id names: blender = a Blender model (its blenderModelId), textureSet = a texture set (its setId). |
| owned | boolean | – | true when the asset is one of the user's own, not one you created: it needs asset:delete:owned. Only when the user asked to delete it. Leave out for an asset you created. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| cascadeImageIds | array | – | – |
| confirmRequired | boolean | yes | – |
| deleted | boolean | yes | – |
| environment | string | yes | – |
| id | string | yes | – |
| kind | string | yes | – |
| meshesDeleted | number | – | – |
| meshesKept | number | – | – |
| name | string|null | yes | – |
| previewProject | – | – | – |
| projects | array | yes | – |
| referenceCount | number|null | yes | – |
| releases | array | – | – |
| releasesDeleted | array | – | – |
| releasesMarkedIncomplete | array | – | – |
| size | number|null | – | – |
No examples provided.
r3_asset_share Show or hide one of your images, audio files or texture sets in the Community ~259
Show one of the user's OWN images, audio files or texture sets in the R3 Community, or hide it again: the same switch as 'Show in Community' in the R3 app. public: true makes it visible to every R3 member, who can then find it in the Community lists and use it in their own projects; public: false hides it again (call this tool again with public: false to undo). A texture set is shown or hidden with all its maps; a map of a texture set is not switched on its own (switch the set). Use it only when the user asked to share or hide that asset. Ids come from r3_image_list, r3_audio_list and r3_texture_read / r3_texture_create. Needs asset:share. Videos, Blender models and projects are shared in the R3 app.
| Name | Type | Req | Description |
|---|---|---|---|
| id | string | yes | The asset's id, as the list tools report it. |
| public | boolean | yes | true shows it in the Community (every R3 member can see and use it); false hides it again. |
| type | string | yes | What the id names: image, audio (an audio file) or textureSet (a texture set, its setId). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| environment | string | yes | – |
| id | string | yes | – |
| name | string|null | yes | – |
| public | boolean | yes | – |
| type | string | yes | – |
No examples provided.
r3_assets_unused List unused images, models and audio ~254
The account's images, models and audio files that nothing uses — no project, no entity template, no other user's project — with each one's size and date, biggest first, and the total. The same report as the R3 app's Settings → Storage, by the same rules the deletes use. A model lists the images only it uses, an HDR / EXR image its preview image: they go with it. A texture set nothing uses is one item (kind textureSet) with its maps under it; a map of a set is never listed on its own. Blender (.blend) files are not part of it: they are the user's to keep or delete, listed by r3_blender_list. Pass `kind` to see one kind, `limit` for more rows (the totals always cover everything). It only reads; r3_asset_delete deletes an item by its kind and id (a plan, then a confirm), for what the user asked to remove. The scan can take several seconds.
| Name | Type | Req | Description |
|---|---|---|---|
| kind | string | – | Only this kind: image, model, audio or textureSet. |
| limit | integer | – | How many items to list (default 50, at most 200). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| environment | string | yes | – |
| items | array | yes | – |
| shown | number | yes | – |
| totals | object | yes | – |
No examples provided.
r3_audio_list List the audio files on this R3 account ~273
List the audio files this R3 account owns, with the URL to use for each one. That URL is exactly what R3ComponentAudio.url takes, so this is how you find a sound the account already has — a footstep, a drum hit, a music bed — and then point an Audio component at it with r3_component_set. Search by name rather than paging: search: "kick" is matched against every audio file on the account by the server, not against the page you happen to be holding. You cannot upload an audio file from here — a tool argument is JSON, this server cannot read your disk, and audio is megabytes. A person uploads those in the R3 app. r3_sound_create makes a new sound effect or seamless loop from a recipe (rendered by R3, listed here afterwards); R3's music components (Transport, Beat, Pattern, DrumGrid) synthesise music that stays editable and in tempo.
| Name | Type | Req | Description |
|---|---|---|---|
| cursor | string | – | Continue a previous listing: pass the nextCursor it returned, unchanged. Omit for the first page. |
| limit | integer | – | How many to return (default 25, max 100). |
| search | string | – | Match this text against the file names, on the server. Omit for everything. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| audios | array | yes | – |
| hasMore | boolean | yes | – |
| limit | number | yes | – |
| nextCursor | string|null | yes | – |
| search | string|null | yes | – |
No examples provided.
r3_blender_copies Find duplicate copies of a Blender model in a project ~139
Does this project hold the same Blender model more than once (two whole imports side by side)? Answers "N copies of <model>: … Keep one?" with each copy (its Entity, entityRef, part count) and its delete plan. It changes nothing: which copy stays is the user's choice, and the other is removed with r3_editor_delete({projectId, componentIds, confirm: {mode: "bundle", planToken}}) once they have chosen.
| Name | Type | Req | Description |
|---|---|---|---|
| blenderModelId | string | yes | The Blender model, from r3_blender_list. |
| projectId | string | yes | The project to look in. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| blenderModelId | string | yes | – |
| copies | array | yes | – |
| environment | string | yes | – |
| modelName | string|null | yes | – |
| projectId | string | yes | – |
| question | string|null | yes | – |
No examples provided.
r3_blender_download Download one of your Blender models as a .blend ~177
Get a .blend of one of the user's OWN Blender models (from r3_blender_list) to open in Blender: the current version, or the version fileId names. Answers a one-time download ticket (5 minutes) and the download endpoint (URL, method, Authorization header); the response body is the file. When the current version is the copy R3 made by applying modifiers, the answer says so and names the original upload, with its modifiers intact, to download instead. Another user's model cannot be downloaded.
| Name | Type | Req | Description |
|---|---|---|---|
| blenderModelId | string | yes | The Blender model, from r3_blender_list (one of the user's own). |
| fileId | string | – | A version's sha256 (from this tool's answer, e.g. the original of an applied copy). Leave out for the current version. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| blenderModelId | string | yes | – |
| body | string | yes | – |
| bytes | number | yes | – |
| environment | string | yes | – |
| expiresAt | string | yes | – |
| fileId | string | yes | – |
| fileName | string | yes | – |
| headers | object | yes | – |
| maxBytes | number | yes | – |
| method | string | yes | – |
| originalFileId | string|null | yes | – |
| ticket | string | yes | – |
| url | string | yes | – |
| version | number|null | yes | – |
No examples provided.
r3_blender_draft_push Show your work in progress on the Blender page ~200
While you build a model in Blender for the user, push an APPLIED copy of your working .blend after each step (body done, wheels done, materials done): the user's Blender page shows it as a 3D preview they can orbit. A draft is not a saved version: one draft per user, the newest wins, kept 24 hours. Answers a one-time ticket (5 minutes) and the draft endpoint (URL, method, Authorization header; the body is the file's raw bytes); sending it needs file access where the .blend is. The draft is a copy with its modifiers applied and its textures packed; r3_blender_upload saves the real file once the user is happy. Up to 20 MB. Needs the blender:upload right.
| Name | Type | Req | Description |
|---|---|---|---|
| model | string | yes | The model's name, as the page shows it: "monster-truck" (the same name you will save it under). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| body | string | yes | – |
| environment | string | yes | – |
| expiresAt | string | yes | – |
| headers | object | yes | – |
| maxBytes | number | yes | – |
| method | string | yes | – |
| model | string | yes | – |
| ticket | string | yes | – |
| url | string | yes | – |
No examples provided.
r3_blender_import Import a Blender model into an R3 project ~414
Add a Blender model the account has uploaded (its current version) to an existing R3 project — the whole model, or only the meshes you name. Every Blender mesh object becomes an R3 Mesh keeping its own position, rotation, scale and origin, with its geometry, materials and textures wired up; objects sharing one mesh share one geometry; a Blender Empty becomes an R3 Empty, the objects under it keeping their place relative to it. A whole model is one Entity whose parts hang from a `<model>_root` Empty: move that Empty to move the model. Reports each object's entityRef so you can reach it with r3_component_set or from code. Use r3_blender_list to choose a model. Pass reparse: true to re-read the .blend itself when the import warns that its meshes were placed at the origin; other projects keep the meshes they use (r3_blender_update moves them, if the user wants). Pass camera to frame the model in a NEW project (its first model: refused on a project that already holds a mesh).
| Name | Type | Req | Description |
|---|---|---|---|
| blenderModelId | string | yes | Which Blender model, from r3_blender_list. |
| camera | object | – | Set the project camera to frame the model, for a NEW project's first model only: refused, with nothing added, when the project already holds a mesh (its camera is then the user's). Omit it to keep th… |
| meshes | array | – | Import only these meshes (by mesh name), each with its own placement. Default: the whole model. |
| projectId | string | yes | The project to import into, from r3_project_create. |
| reparse | boolean | – | Re-read the .blend instead of using what was stored at upload time. This is R3's Re-parse button. Slow (it re-parses and re-optimises every texture) but it is the only way to recover object placement… |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| appliedInBlender | – | – | – |
| blenderModelId | string | yes | – |
| cameraControlsAdded | array | – | – |
| cameraSet | boolean | – | – |
| componentCount | number | – | – |
| entity | – | – | – |
| fileId | string|null | yes | – |
| filename | string|null | yes | – |
| materialCount | number | – | – |
| meshCount | number | – | – |
| objects | array | – | – |
| projectId | string | yes | – |
| reparsed | boolean | yes | – |
| root | – | – | – |
| status | string | yes | – |
| textureCount | number | – | – |
| warnings | array | – | – |
No examples provided.
r3_blender_list List the Blender models on this R3 account ~277
List the Blender models this R3 account has uploaded — one per .blend file name, the current version of each — with what it will produce: how many Blender objects land in the scene, how many meshes and materials, its screenshot, and any PENDING update (a re-upload that changes meshes the user's projects use, waiting for r3_blender_update). An upload with unapplied modifiers is applied in Blender by R3 itself; a model says what was applied, per object. community: true lists everyone's public models instead. Add one (or some of its meshes) to a project with r3_blender_import. A .blend gets here two ways: a person uploads it in the R3 app (Models -> Upload .blend), or you save one you built in Blender with r3_blender_upload (a one-time ticket and an upload endpoint: never a tool argument).
| Name | Type | Req | Description |
|---|---|---|---|
| community | boolean | – | true lists the PUBLIC models of every user (the Community tab) instead of this account's own. |
| cursor | string | – | Continue a list: the nextCursor of the previous page. |
| limit | integer | – | How many models to return (default 25, max 100). |
| search | string | – | Only models whose name matches, searched across every model (not the page in hand). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| filter | string | yes | – |
| hasMore | boolean | yes | – |
| models | array | yes | – |
| nextCursor | string|null | yes | – |
No examples provided.
r3_blender_update Update projects to a Blender model's current version ~445
A re-uploaded Blender model is always its latest version; the user's projects keep the meshes they have until asked. This tool asks and answers that one question: choice update_projects without reportId shows which projects and entity templates use an older version (or an earlier import of the same file) and what changed; with the reportId and the projectIds / entityTemplateIds the user chose, it moves just those to the current version. Which ones move is the user's choice; "not now" means no call. Parts are matched by Blender OBJECT name: each takes its new mesh and placement; when only its shape changed it keeps the project's materials and the edits made to them in R3; when its materials changed in Blender it takes Blender's materials where the import's own material is unchanged, while a material made or changed in R3 is kept unless the user chooses Blender's (replaceMaterials, from the plan's materialChoices), and a material made in R3 is never deleted. Objects new in the model are added under the copy's root Empty, parts the model no longer has are kept (removeGone: true removes them). Published releases never change. The plan only reads; moving projects needs the blender:update right.
| Name | Type | Req | Description |
|---|---|---|---|
| blenderModelId | string | yes | The Blender model, from r3_blender_list. |
| choice | string | yes | Required: update_projects. There is no default. |
| entityTemplateIds | array | – | With a reportId: the entity templates to update. |
| projectIds | array | – | With a reportId: the projects to update. |
| removeGone | boolean | – | Also remove the parts the current version no longer has (default false: they are kept and listed). Plan with it and answer that report with it. |
| replaceMaterials | array | – | With a reportId: material ids from the plan's materialChoices (materials made or changed in R3, kept by default) the user chose to replace with Blender's. Default none: R3's materials are kept. |
| reportId | string | – | The report being answered, from the plan (this tool without reportId). Leave it out to plan. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| behindProjectIds | array | – | – |
| blenderModelId | string | yes | – |
| choice | string | yes | – |
| currentFileId | string|null | yes | – |
| duplicates | array | – | – |
| environment | string | yes | – |
| freedRows | number|null | yes | – |
| kept | array | yes | – |
| live | boolean | – | – |
| materialChoices | array | – | – |
| projectParts | array | – | – |
| repointed | array | yes | – |
| reportId | string | – | – |
| versionNumber | number | – | – |
No examples provided.
r3_blender_upload Save a .blend as a version of a Blender model ~277
Save a .blend you made or changed (in Blender, through a Blender MCP) as the newest version of one of the user's Blender models: the model is named by the file name, so "monster-truck.blend" adds a version to the user's monster-truck model, or makes the model. A tool argument cannot carry the file, so this answers a one-time upload ticket (5 minutes) and the upload endpoint: its URL, method, Authorization header and that the body is the file's raw bytes. Sending it needs file access where the .blend is (the R3 sandbox, or the user's own machine); without it, the user uploads on the Blender page. The endpoint's answer names the model, the version number and its sha256; R3 applies unapplied modifiers itself, and when the user's projects use an older version it asks one question (answer it with r3_blender_update). Then r3_blender_import puts the model in a project. Textures travel only when packed (bpy.ops.file.pack_all()). Up to 100 MB. Needs the blender:upload right.
| Name | Type | Req | Description |
|---|---|---|---|
| fileName | string | yes | The .blend file name, which names the model: "monster-truck.blend" (".blend" is added when missing). A name only, no folder. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| body | string | yes | – |
| environment | string | yes | – |
| expiresAt | string | yes | – |
| fileName | string | yes | – |
| headers | object | yes | – |
| maxBytes | number | yes | – |
| method | string | yes | – |
| ticket | string | yes | – |
| url | string | yes | – |
No examples provided.
r3_bug_report File a bug report about R3 ~343
Send R3 a bug report about R3 ITSELF (the engine, the editor, these tools), in the user's name and naming you (your agent key or app connection). It becomes one of the user's support issues, which only the user and R3's team see. Use it when R3 does something its documentation says it should not, after checking the docs (r3_docs_read) — not for problems in the user's own project, and not for feature wishes (those are r3_roadmap_suggest). r3_bug_reports_mine lists what the user already reported, so a fault is not filed twice. A useful report says what was done (the tool calls), what was expected and what happened, with what was MEASURED in `measurements` — the values read back with r3_component_read or r3_editor_inspect, the exact error text. `projectId` (when it happens in a project) lets R3 open it. The answer carries the report's link for the user. Rate-limited per agent key or app connection (a few an hour). Needs the support:write right.
| Name | Type | Req | Description |
|---|---|---|---|
| description | string | yes | What you did, what you expected, what happened instead. |
| measurements | string | – | What you measured: values read back, exact error text, timings. Plain text. |
| projectId | string | – | The project it happens in (r3_project_list), if any. Only a project the user can open. |
| title | string | yes | One line naming the fault, e.g. "Cuboid collider with halfExtents.z = 0 inverts the contact normal". |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| issueId | string | yes | – |
| projectId | string|null | yes | – |
| status | string | yes | – |
| title | string | yes | – |
| url | string | yes | – |
No examples provided.
r3_bug_reports_mine List the bug reports the user has filed ~106
The user's own support issues — bug reports and questions they or their agents (agent keys and app connections) sent R3 — newest first, with status, date, the project and whether an agent filed it (and which), each with its link. It shows whether a fault is already reported, before r3_bug_report would file it twice. Needs the support:write right.
| Name | Type | Req | Description |
|---|---|---|---|
| limit | integer | – | How many to return, newest first (default 20). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| issues | array | yes | – |
| total | number | yes | – |
No examples provided.
r3_code_read Read the code and event binding of R3 Code components ~455
Read back what a Code component actually contains — its full JavaScript source and the event it runs on. r3_code_write reports what it wrote; this reports what is there. Use it before replacing a script, so an edit builds on the real source instead of overwriting it blind, and after a write whose reply you did not see, to confirm what landed. Call it with NO entityRefs to get every Code component in the project — that is how you find a script again when you never received the entityRef for it. Each script comes back with its codeVersion: pass it as r3_code_write's baseVersion. The event comes back as a NAME (RENDER_BEFORE, KEY_DOWN, …), which is the same form r3_code_write takes, so you can hand it straight back. r3_component_read reports the same field as a raw number and omits any source over 2000 characters; this is the tool for reading code. Name ONE entityRef to read it in pages (offset / length): every script comes back with totalChars, the sha256 of its WHOLE source, and truncated / nextOffset. When truncated is true you have only part of it: read on from nextOffset, join the pages and check the sha256 before writing it back. Reading several, a body longer than one page is left out WHOLE and named with its size — never cut silently, because a partial script written back would delete the rest of it. An entityRef that does not exist is named and the others are still read.
| Name | Type | Req | Description |
|---|---|---|---|
| entityRefs | array | – | Which Code components to read, e.g. ["spin"]. Omit for ALL of them. Up to 50 per call. |
| length | integer | – | How many characters to read at most (default and maximum 60000; 8000 in an R3 Builder session). A page ends on a whole line when it can. |
| offset | integer | – | Where to start reading, in characters (default 0). For the next page pass the nextOffset of the previous answer. |
| projectId | string | yes | The project to read from, from r3_project_create or r3_project_load. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| components | array | yes | – |
| projectId | string | yes | – |
| projectName | string|null | yes | – |
| unreadable | array | yes | – |
No examples provided.
r3_code_write Write code that runs in an R3 project ~894
Add or replace a Code component: JavaScript a project runs. It runs on ONE event you choose (`event`): RENDER_BEFORE (default) every frame, for movement (held keys: this.keysPressed); KEY_DOWN / KEY_UP once per press / release; MOUSE_CLICK, SWIPE, MUSIC_BEAT, PHYSICS_COLLISION once per click, swipe, beat, collision. A script per event: r3_docs_read("HOWTO_CODE_EVENTS.md"). Kept at once in the live copy every open R3 tab runs (r3_editor_code: the same tool). r3_project_load names the entityRefs; r3_docs_read("CODE_PATTERNS.md") is the API (a guessed one silently does nothing). The code is a function body: ```js this.update = function(data) { this['mesh_cube_1'].rotation.y += data.delta; }; return {update: this.update}; // REQUIRED (also start, dispose) — nothing runs without it ``` Reach a component as `this['<entityRef>']`, infrastructure with this.getProject() / getScene() / getCamera() / getWorld() / getAudioListener() / getEntityInstance(). Anything else returned is an api other scripts call: `this['<entityRef>'].api.<name>(...)`. Code that creates things (scene objects, timers, listeners) frees them in a returned `dispose`: R3 calls it on re-save, Stop and delete. No getComponentByRef, this.state, this.getSibling or R3 global: code using them is REFUSED. EVENTS (by NAME; default RENDER_BEFORE): RENDER_BEFORE, RENDER_AFTER, KEY_DOWN, KEY_UP, MOUSE_DOWN, MOUSE_UP, MOUSE_MOVE, MOUSE_WHEEL, MOUSE_CLICK, TOUCH_START, TOUCH_MOVE, TOUCH_END, TOUCH_CANCEL, SWIPE, MUSIC_BEAT, MUSIC_BAR, SONG_CUE, PHYSICS_COLLISION. Their `data` (mouse: nested, -1..+1): the how-to's table. Pass entityRef + baseVersion (r3_code_read's codeVersion) to REPLACE a script; name (or neither) to create one. Never merged: an older one is refused with the current code. PATCH a long script: replacements [{old, new, count?}] + baseSha256 (r3_code_read's), no code; all or nothing, each old found exactly count (default 1) times; ≤ 200 replacements, 60000 chars.
| Name | Type | Req | Description |
|---|---|---|---|
| baseSha256 | string | – | With replacements only: the sha256 r3_code_read answered for the script (of its WHOLE code) at baseVersion. |
| baseVersion | integer | – | REQUIRED with entityRef: the codeVersion r3_code_read answered for that script. A write on an older version is refused with the current code, never merged. Leave it out when creating. |
| code | string | – | The JavaScript. A function body that returns {start, update} — see the description. Omit it when you send replacements. |
| componentId | string | – | Or the id of the EXISTING Code component to replace (instead of entityRef). |
| entityRef | string | – | The entityRef of an EXISTING Code component to replace, as reported by r3_project_load. Cannot be combined with name. |
| event | string | – | Which event runs it, by NAME: RENDER_BEFORE, RENDER_AFTER, KEY_DOWN, KEY_UP, MOUSE_DOWN, MOUSE_UP, MOUSE_MOVE, MOUSE_WHEEL, MOUSE_CLICK, TOUCH_START, TOUCH_MOVE, TOUCH_END, TOUCH_CANCEL, SWIPE, MUSIC… |
| name | string | – | Label for a NEW Code component, e.g. "Spin". Cannot be combined with entityRef. |
| projectId | string | yes | The project to write into, from r3_project_create or r3_project_load. |
| replacements | array | – | Instead of code: exact replacements on the script at baseVersion, in order, all or nothing (at most 200, 60000 characters of old + new in all). |
| Name | Type | Req | Description |
|---|---|---|---|
| action | string | yes | – |
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| codeVersion | integer | yes | – |
| componentId | string | yes | – |
| entityRef | string | yes | – |
| event | string | yes | – |
| fileName | string | yes | – |
| imports | array | – | – |
| name | string | yes | – |
| occurrences | – | yes | – |
| projectId | string | yes | – |
| replacements | – | yes | – |
| seq | – | – | – |
| sha256 | string | yes | – |
| totalChars | integer | yes | – |
| warnings | array | – | – |
No examples provided.
r3_component_add Add ANY component to a saved R3 project ~507
Add one component of ANY registered type — a collider, a rigid body, a joint, a light, a material, a geometry … — to a saved R3 project, with its values and references set before it is created. Kept at once in the project's live copy: every R3 tab with the project open shows it now, no editor needed and nothing to save (r3_editor_add_component is the same tool). Use r3_factory_run for a ready-made object (a cube, a physics sphere …). `type` is the short type name, e.g. "ColliderCuboid" (an unknown type is refused with the list). `values` take r3_component_set's shapes: a plain value {"friction": 0.2}; a value component as the keys to set {"halfExtents": {"x": 2, "y": 0.5, "z": 1}}; a reference as the target's entityRef or id {"rigidBody": "rigidbody_dynamic_1"} — a component of THIS project, of a type the slot accepts; "name" names it (a name already used is kept as its label, and its entityRef gets a suffix, e.g. hull_2: the answer says so); "entity" puts it in an Entity (its value components join too). A type that cannot exist without an option (an Audio / AudioPositional / Texture's "url") takes it in values too, and is refused without it. A Code component's code is r3_code_write's. A VehicleWheel added with a mesh is measured from it (Configure from Mesh: radius, connectionPoint, axle), and a ColliderCuboid / Sphere / Cylinder / Capsule with a mesh (its own or its rigid body's) is fitted to it (Fit to Mesh: size, and position on the mesh's centre), only the values you did not pass; the answer says what (PHYSICS_USAGE.md → Vehicles).
| Name | Type | Req | Description |
|---|---|---|---|
| projectId | string | yes | The project to add to (r3_project_list). |
| type | string | yes | The component type, e.g. "ColliderCuboid", "RigidBodyFixed", "LightPoint". |
| values | object | – | Its values, e.g. {"rigidBody": "rigidbody_dynamic_1", "halfExtents": {"x": 2}, "friction": 0.2}. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| componentId | string | yes | – |
| created | number | yes | – |
| entityRef | string|null | yes | – |
| physics | array | yes | – |
| projectId | string | yes | – |
| saved | boolean | yes | – |
| seq | – | – | – |
| type | string | yes | – |
No examples provided.
r3_component_bounds Measure how big a mesh or model is ~279
The size and place of meshes and imported models, in R3 units (metres): the world-space box around each (min, max, size, centre), the mesh's scale, rotation and parents applied. Name a Mesh, or the root Empty of a Blender model (r3_blender_import makes one, named "<model>_root"), and the box covers it and every mesh under it, with each part listed. Use it before scaling or placing anything relative to its size: to stand a model on the ground, move it so its min.y is the ground height (r3_terrain_height on a terrain); to make it 2 m long, scale by 2 / size. Each answer also gives the project camera's position and lookAt that show it whole on any screen, a phone held upright included: use it to frame a model. Several in one call (up to 20). Works with no R3 tab open, from the project as it is kept; read-only.
| Name | Type | Req | Description |
|---|---|---|---|
| entityRefs | array | yes | Meshes or model root Empties to measure, by entityRef (r3_project_load lists them) or id, e.g. ["truck_root"]. Up to 20. |
| projectId | string | yes | The project, from r3_project_list or r3_project_create. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| components | array | yes | – |
| projectId | string | yes | – |
| projectName | string|null | yes | – |
| unreadable | array | yes | – |
No examples provided.
r3_component_read Read the property values of R3 components ~439
Read back what components actually hold — a material's colour and roughness, a mesh's position, which material it uses. r3_project_load gives you the names; this gives you the values behind them. Pass SEVERAL entityRefs in one call — telling six similarly-named materials apart is the case this exists for, and it is one round trip rather than six. One component: entityRef: "mesh_cube_1" (r3_component_set's name for it) works too. What comes back is exactly what r3_component_set takes, so read → change → write needs no translation: value components are expanded inline ({"position": {"x": 0, "y": 2, "z": 0}}), references are the target's entityRef as a string (or null when empty), and lists of references are arrays of entityRefs. Use it before any relative change ("20% brighter", "move it up by one"), and after a write to confirm what landed. Bulk data — mesh vertex arrays and anything else megabyte-sized — is NOT returned. Every property left out is listed with the reason, so an omission is never silent. Each answer says which entity the component belongs to (entity: {entityRef, id}, or null). An id works in place of an entityRef; an entityRef several components hold (the same name in different entities) is not guessed: it is reported with each holder's id. An entityRef that does not exist is reported by name and the rest are still read. Value components (a Vector3, a Color) are not read on their own: ask for the component that owns them.
| Name | Type | Req | Description |
|---|---|---|---|
| entityRef | string | – | Or one component, e.g. "mesh_cube_1". |
| entityRefs | array | – | Which components to read, as reported by r3_project_load: an ARRAY, e.g. ["mesh_cube_1", "material_cube_1"]. Up to 50 per call. |
| projectId | string | yes | The project the components are in, from r3_project_create or r3_project_load. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| components | array | yes | – |
| projectId | string | yes | – |
| projectName | string|null | yes | – |
| unreadable | array | yes | – |
No examples provided.
r3_component_rename Rename a component (its entityRef) ~402
Rename a component: its entityRef, the name scripts address it by (this['name'] / this.name), and its display name when that is where the entityRef came from. Name it by entityRef or componentId. Without planId it only PLANS and changes nothing: it checks the new name (refused: not an entityRef R3 makes, a reserved name, or taken in the same namespace: inside the same Entity, or at project level, naming the holder), and lists every Code script whose this['old'] / this.old finds THIS component (a script inside an Entity finds its own member first), with codeVersion and line numbers, plus scripts where the name is ambiguous and scripts that already read the new name. With that planId it applies the plan as ONE undoable change every open R3 tab sees. updateCode: true also rewrites those scripts in the same change (each on the version planned: one changed since refuses the whole rename, never merged); without it they keep the old name and silently find nothing. If anything changed since the plan, nothing is written and the new plan is returned.
| Name | Type | Req | Description |
|---|---|---|---|
| componentId | string | – | The component to rename, by its id (needed when several share the entityRef). |
| entityRef | string | – | The component to rename, by its entityRef (or its id). Or use componentId. |
| newRef | string | yes | The new entityRef: lower case, _ instead of spaces, @ and dots (e.g. "arena_ground"). |
| planId | string | – | The plan being applied, from this tool without planId (same component and newRef). Leave it out to plan. |
| projectId | string | yes | The project, from r3_project_load. |
| updateCode | boolean | – | With planId: also rewrite this['old'] / this.old in the listed scripts, in the same change. Default false. |
| Name | Type | Req | Description |
|---|---|---|---|
| ambiguousScripts | array | yes | – |
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| applicable | boolean | yes | – |
| applied | boolean | yes | – |
| component | object | yes | – |
| environment | string | yes | – |
| modulesScanned | boolean | yes | – |
| nameChange | – | yes | – |
| newNameReaders | array | yes | – |
| newRef | string | yes | – |
| notPatched | array | yes | – |
| oldRef | string | yes | – |
| patched | array | yes | – |
| planId | string | yes | – |
| projectId | string | yes | – |
| refusals | array | yes | – |
| scripts | array | yes | – |
| seq | – | yes | – |
| template | – | yes | – |
| templateCopies | array | yes | – |
| updateCode | boolean | yes | – |
No examples provided.
r3_component_set Change a property on an R3 component ~1,003
Set one or more properties on a component that already exists in a project — a material's roughness, a mesh's position, which material a mesh uses. Kept at once in the project's live copy: every R3 tab with the project open shows it now, no editor needed and nothing to save (r3_editor_set is the same tool). The answer says what was kept and its seq. Components are addressed by entityRef (r3_project_load lists them); one several components share is refused with them: pass componentId (or the id as a reference). {"entity": "<an Entity's entityRef>"} puts it in that Entity (with its position, colours …); null takes it out. Properties come in four shapes and the value you pass differs: • plain value — {"roughness": 0.25}, {"visible": false} • value component (position / rotation / scale / color / size …) — pass an OBJECT of just the keys to change: {"position": {"y": 2}} leaves x and z alone. Vector3 has x/y/z, Color has r/g/b/a (0..1), Scalar has value. An EMPTY slot (a rigid body's centerOfMass is empty until set) is made from the fields given: {"centerOfMass": {"z": -0.3, "mass": 8}}; the fields not named keep their defaults. Rotations are in RADIANS, not degrees: a quarter turn is Math.PI/2 ≈ 1.5708 (the R3 Inspector shows rotations in degrees: 90°). • reference to another component — pass that component's entityRef as a string: {"material": "material_cube_1"}, or null to clear it. • list of references (a mesh's materials) — pass an ARRAY of entityRefs: {"materials": ["mat_a", "mat_b"]}. The array REPLACES the whole list, in the order given — it does NOT append, so name every component you want in it; [] empties the list. The result reports the list before and after, so check that against what you intended. Several properties in one call apply all-or-nothing. An unknown property is refused with the valid names (a typo never corrupts the project). Rename (name / entityRef) with r3_component_rename. Giving a VehicleWheel its mesh or vehicle measures the…
| Name | Type | Req | Description |
|---|---|---|---|
| componentId | string | – | Or its id (r3_project_load) — exactly one of entityRef / componentId. |
| entityRef | string | – | Which component to change, as reported by r3_project_load, e.g. "mesh_cube_1". |
| projectId | string | yes | The project the component is in, from r3_project_create or r3_project_load. |
| properties | object | yes | What to set: {"roughness": 0.25, "position": {"y": 2}, "material": "material_cube_1"}. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| changes | array | yes | – |
| componentId | string | yes | – |
| created | array | yes | – |
| entityRef | string | yes | – |
| keptWorld | object | – | – |
| landed | array | – | – |
| physics | array | yes | – |
| projectId | string | yes | – |
| seq | – | – | – |
| type | string | yes | – |
No examples provided.
r3_docs_list List the R3 documentation ~76
Every R3 document this server can give you, IN READING ORDER, with what each one covers and how big it is. R3's documentation begins here: no model has R3 in its pretraining, and code written from tool descriptions alone saves cleanly and then silently does nothing. r3_docs_read returns one.
Input schema present but exposes no named parameters.
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| apiIndexPath | string | yes | – |
| docs | array | yes | – |
No examples provided.
r3_docs_read Read one R3 document ~234
Return one R3 document in full, or one API symbol. Paths come from r3_docs_list: r3_docs_read("CODE_PATTERNS.md") how a Code component is written (reading order 1) r3_docs_read("PHYSICS_USAGE.md") colliders, rigid bodies, joints r3_docs_read("HOWTO_PHYSICS.md") one task step by step (how-tos: r3_docs_list) r3_docs_read("api") index of every component, system and event r3_docs_read("api/component/Mesh") one component r3_docs_read("api/system/PhysicsRapier") one system r3_docs_read("api/event/KEY_DOWN") one event, and the data it delivers The API reference is only ever read one symbol at a time — the whole file is too large to serve.
| Name | Type | Req | Description |
|---|---|---|---|
| path | string | yes | Which document, e.g. "CODE_PATTERNS.md", "api", "api/component/Mesh". From r3_docs_list. |
No output schema declared.
No examples provided.
r3_editor_add_component Add ANY component LIVE to the user's open R3 editor ~354
The same tool as r3_component_add (this is its older name): add one component of ANY registered type — a collider, a rigid body, a joint, a light, a material, a geometry … — with its values and references set before it is created, kept at once in the project's live copy: every R3 tab with the project open shows it now, no editor needed and nothing to save. Use r3_editor_create for a ready-made object. `type` is the short type name, e.g. "ColliderCuboid" (an unknown type is refused with the list). `values` take r3_component_set's shapes: a plain value {"friction": 0.2}; a value component as the keys to set {"halfExtents": {"x": 2, "y": 0.5, "z": 1}}; a reference as the target's entityRef or id {"rigidBody": "rigidbody_dynamic_1"} — a component of THIS project, of a type the slot accepts. Example — swap a rigid body's collider: add a ColliderCuboid with {"rigidBody": …, "halfExtents": …}, then r3_editor_delete the old collider (mode "one").
| Name | Type | Req | Description |
|---|---|---|---|
| projectId | string | yes | The project (r3_project_list). |
| type | string | yes | The component type, e.g. "ColliderCuboid", "RigidBodyFixed", "LightPoint". |
| values | object | – | Its values, e.g. {"rigidBody": "rigidbody_dynamic_1", "halfExtents": {"x": 2}, "friction": 0.2}. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| componentId | string | yes | – |
| created | number | yes | – |
| editor | – | yes | – |
| entityRef | string|null | yes | – |
| live | boolean | yes | – |
| projectId | string | yes | – |
| saved | boolean | yes | – |
| seq | – | – | – |
| type | string | yes | – |
No examples provided.
r3_editor_capture Take a picture of the project in the user's open R3 editor ~290
Capture the canvas of a project open in the user's R3 editor — the scene as it looks NOW, running physics included — and LOOK at it: the answer is an image. Call it after a visual change (r3_editor_set, r3_editor_create, r3_editor_code …) to check what you built, and after r3_editor_run to see it move. By default (`camera: "editor"`) it is what the user sees: the editor's view camera, edit aids included. `camera: "project"` takes it through the project's own camera instead, with no edit aids, as a visitor sees it: use that to check a camera or a camera script. At most 1280 px each way (`width`, default 1024); a JPEG. Needs project:read.
| Name | Type | Req | Description |
|---|---|---|---|
| camera | string | – | Whose view: "editor" (default) = the editor's view camera the user looks through, edit aids included; "project" = the project's own camera, no edit aids, as a visitor sees it (check a camera script w… |
| projectId | string | yes | A project open in the user's editor (r3_whoami lists them). |
| width | number | – | The picture's largest width in px, 64..1280 (default 1024); it keeps the canvas's aspect and is never scaled up. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| bytes | number | yes | – |
| camera | string | yes | – |
| editor | – | yes | – |
| height | number | yes | – |
| mimeType | string | yes | – |
| projectId | string | yes | – |
| width | number | yes | – |
No examples provided.
r3_editor_code Run new code LIVE in the user's open R3 editor ~377
The same tool as r3_code_write (this is its older name): replace a Code component's code, kept at once in the project's live copy: every R3 tab with the project open runs the new code now, no editor needed and nothing to save. Name it with componentId or entityRef and pass baseVersion, the codeVersion r3_code_read answered: code is never merged, so on an older version it is refused with the current code and version (apply your change to that and call again). A long script can be PATCHED instead: replacements + baseSha256, as r3_code_write takes them. The code is a FUNCTION BODY (a top-level `return {start, update};`); read r3_docs_read("CODE_PATTERNS.md") for the R3 APIs. A syntax error or an R3 API that does not exist is refused before anything is kept.
| Name | Type | Req | Description |
|---|---|---|---|
| baseSha256 | string | – | With replacements only: the sha256 r3_code_read answered for the script (of its WHOLE code) at baseVersion. |
| baseVersion | – | – | REQUIRED: the codeVersion r3_editor_inspect gave for this Code component (the version your code builds on); null for one never saved (codeVersion null). |
| code | string | – | The whole new code (a function body). Omit it when you send replacements. |
| componentId | string | – | The Code component's id. |
| entityRef | string | – | Or its entityRef, e.g. "code_spin". |
| projectId | string | yes | The project (r3_project_list). |
| replacements | array | – | Instead of code: exact replacements on the script at baseVersion, in order, all or nothing (at most 200, 60000 characters of old + new in all). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| codeVersion | – | yes | – |
| compiled | boolean | yes | – |
| componentId | string | yes | – |
| editor | – | yes | – |
| entityRef | string|null | yes | – |
| error | string|null | yes | – |
| live | boolean | yes | – |
| name | string|null | yes | – |
| projectId | string | yes | – |
| saved | boolean | yes | – |
| seq | – | – | – |
| sha256 | string | yes | – |
| totalChars | integer | yes | – |
| warnings | array | yes | – |
No examples provided.
r3_editor_create Add an object LIVE to the user's open R3 editor ~229
The same tool as r3_factory_run (this is its older name): add an object — a cube, sphere, light, physics body … — kept at once in the project's live copy: every R3 tab with the project open shows it now, no editor needed and nothing to save. r3_factory_list lists the factories and their params.
| Name | Type | Req | Description |
|---|---|---|---|
| factory | string | yes | Factory name, e.g. "Cube", "Sphere", "PhysicsCube". See r3_factory_list. |
| name | string | – | The name of what you make, e.g. "Leg front left" (r3_factory_run's name). |
| params | object | – | Factory options, e.g. {"color": 16711680, "position": {"y": 2}}. Color is a packed RGB integer. Rotations are in RADIANS, not degrees: a quarter turn is Math.PI/2 ≈ 1.5708 (the R3 Inspector shows rot… |
| projectId | string | yes | The project (r3_project_list). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| created | number | yes | – |
| editor | – | yes | – |
| factory | string | yes | – |
| live | boolean | yes | – |
| projectId | string | yes | – |
| roles | object | yes | – |
| saved | boolean | yes | – |
| seq | – | – | – |
No examples provided.
r3_editor_delete Delete components LIVE in the user's open R3 editor (plan, then confirm) ~323
Delete components from the project the user has OPEN in their R3 editor, exactly as the editor's delete dialog asks it — in TWO calls. First call WITHOUT `confirm`: it returns the plan and changes nothing — the selected components, what only they use (their geometry, material, colliders …), an Entity's members, the modes on offer and a planToken. Then call again with the same componentIds and confirm: {"mode": …, "planToken": …}: "one" deletes only the selection, "with_dependents" also what only it uses, "bundle" (one Entity) the entity with its members. The mode is the user's choice, unless they already said. If the project changes in between, the confirm is refused: plan again. Only when the user asked for it. The confirmed delete is kept at once in the project's live copy (every open R3 tab shows it; no editor is needed); R3 keeps what was deleted in the project's trash. The children of a deleted Empty or Mesh keep their place in the world, as the editor's delete does (their numbers become the scene's; the answer lists them). A project itself is never deleted here. Needs the editor:dispose capability.
| Name | Type | Req | Description |
|---|---|---|---|
| componentIds | array | yes | The ids of the components to delete (r3_editor_inspect / r3_project_load). |
| confirm | object | – | Leave out to get the plan; give it to delete what the plan named. |
| projectId | string | yes | The project (r3_project_list). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| confirmed | boolean | yes | – |
| dependents | array | yes | – |
| disposed | array | yes | – |
| editor | – | yes | – |
| live | boolean | yes | – |
| members | array | yes | – |
| mode | string|null | yes | – |
| modes | array | yes | – |
| planToken | string|null | yes | – |
| projectId | string | yes | – |
| saved | boolean | yes | – |
| selected | array | yes | – |
| seq | – | – | – |
No examples provided.
r3_editor_inspect Read live component values from the user's open R3 editor ~372
Read what components hold RIGHT NOW in an R3 tab that has the project open: runtime values included — a moving body's position while the project runs. Use it to debug live behaviour. It needs a tab with the project open (r3_editor_open opens one). r3_component_read reads the kept values without a tab. Select with any of: `ids` (component ids from r3_project_load), `type` ("Mesh", "RigidBodyDynamic", …) and `entityRef`; they combine. `fields` keeps only the properties you name (e.g. ["position", "rotation"]). At most 50 components per call; a larger answer is refused with how to narrow it. Values are in r3_component_read's shape (value components inline, references as entityRefs), so they can be handed to r3_component_set. The project must be open in the editor (r3_editor_open opens it). Read-only: nothing is changed. Script problems: `scriptDiagnostics` lists the recent errors and warnings of the Code components read (compile / start / runtime errors, warnings such as AmbiguousEntityRefError), each with its name, message, count and last time, since the last Run. Inspect type "Code" after r3_editor_run to see whether your scripts run cleanly; null means the tab does not report them.
| Name | Type | Req | Description |
|---|---|---|---|
| entityRef | string | – | Only the component with this entityRef. |
| fields | array | – | Only these properties, e.g. ["position", "rotation"]. |
| ids | array | – | Component ids to read. |
| projectId | string | yes | The project, open in the user's editor. |
| type | string | – | Only components of this type, e.g. "Mesh" or "R3ComponentMesh". |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| components | array | yes | – |
| editor | – | yes | – |
| fieldsNotFound | array | yes | – |
| live | boolean | yes | – |
| projectId | string | yes | – |
| projectName | string|null | yes | – |
| scriptDiagnostics | – | yes | – |
| unreadable | array | yes | – |
No examples provided.
r3_editor_open Show a project in the user's open R3 editor ~204
Open a project in the R3 editor the user has open in their browser (their most recently used R3 tab) and focus it, so they see it. A project already open there is just focused (never reloaded). It is how the user gets to see the project you are working on, and how a tab gets the project r3_editor_inspect reads. To show the project to the user right here in the chat, use r3_preview_create instead (a playable card in the chat where the client shows MCP Apps, else a link). The same access rules as opening it by hand: the user's own projects, their teams', and public ones. The editor holds up to 4 projects; at that limit it refuses. Nothing is saved or changed. When no editor is open, the answer says where the user can open one.
| Name | Type | Req | Description |
|---|---|---|---|
| projectId | string | yes | The project to open, from r3_project_list or r3_project_create. |
| Name | Type | Req | Description |
|---|---|---|---|
| alreadyOpen | boolean | yes | – |
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| editor | – | yes | – |
| focused | boolean | yes | – |
| opened | boolean | yes | – |
| projectId | string | yes | – |
| projectName | string|null | yes | – |
No examples provided.
r3_editor_run Run the project in the user's open R3 editor ~213
Press the Run buttons of the user's open R3 editor: start physics, code, audio and music (all of them, or the ones named in `systems`), exactly as the user's own click does — a paused simulation resumes. Use it to TEST what you built: run it, then look with r3_editor_capture and read live values with r3_editor_inspect. Changes nothing saved. It shows the named project first (focused, scrolled into view): the systems start for every project open in that tab, but only a project on screen moves. The scene only moves while that tab is in front: a browser pauses a background tab, and the answer then says the scene is NOT advancing. Needs the editor:run right.
| Name | Type | Req | Description |
|---|---|---|---|
| projectId | string | yes | A project open in the user's editor (r3_whoami lists them). |
| systems | array | – | Which of the editor's Run / Stop buttons: physics, code, audio, music. Leave out for all of them. |
| Name | Type | Req | Description |
|---|---|---|---|
| action | string | yes | – |
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| editor | – | yes | – |
| projectId | string | yes | – |
| shown | boolean | – | – |
| systems | array | yes | – |
No examples provided.
r3_editor_set Change a property LIVE in the user's open R3 editor ~355
The same tool as r3_component_set (this is its older name): change properties of a component, kept at once in the project's live copy: every R3 tab with the project open shows it now, no editor needed and nothing to save. Name the component with componentId or entityRef. `values` take r3_component_set's shapes: a plain value {"roughness": 0.25}; a value component as an object of the keys to change {"position": {"y": 2}} (an EMPTY slot, e.g. a rigid body's centerOfMass, is made from the fields given: {"centerOfMass": {"z": -0.3, "mass": 8}}); a reference as the target's entityRef {"material": "material_cube_1"} (null clears it); a list of references as an array of entityRefs (it REPLACES the list). All-or-nothing: an unknown name is refused with the valid ones. A Code component's code is r3_editor_code's. Rotations are in RADIANS, not degrees: a quarter turn is Math.PI/2 ≈ 1.5708 (the R3 Inspector shows rotations in degrees: 90°).
| Name | Type | Req | Description |
|---|---|---|---|
| componentId | string | – | The component's id (r3_project_load / r3_editor_inspect). |
| entityRef | string | – | Or its entityRef, e.g. "mesh_cube_1". |
| projectId | string | yes | The project (r3_project_list). |
| values | object | yes | What to set: {"visible": false, "position": {"y": 2}, "material": "material_cube_1"}. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| changes | array | yes | – |
| componentId | string | yes | – |
| editor | – | yes | – |
| entityRef | string|null | yes | – |
| landed | array | yes | – |
| live | boolean | yes | – |
| projectId | string | yes | – |
| saved | boolean | yes | – |
| seq | – | – | – |
| type | string | yes | – |
No examples provided.
r3_editor_set_preview_image Set the project's share image from the user's open R3 editor ~188
Use what the project's camera shows NOW, in the user's open R3 editor, as the project's share image: the picture its live page, link previews (social cards) and the showcase show, used when the project is next published (r3_publish_live, or Publish update in the editor). The editor waits until the scene's textures have loaded and takes it without edit aids (no grid, gizmos or outlines), as a visitor sees it; the answer is the picture it set. It is for once the project looks right (r3_editor_capture shows it beforehand), or when the user asks for a new picture. A live page keeps the share image of its live release until that next publish. Needs a tab with the project open and project:write.
| Name | Type | Req | Description |
|---|---|---|---|
| projectId | string | yes | A project open in the user's editor (r3_whoami lists them). |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| bytes | number | yes | – |
| editor | – | yes | – |
| height | number | yes | – |
| mimeType | string | yes | – |
| projectId | string | yes | – |
| saved | boolean | yes | – |
| width | number | yes | – |
No examples provided.
r3_editor_stop Stop the project in the user's open R3 editor ~138
Press the Stop buttons of the user's open R3 editor: stop physics, code, audio and music (all of them, or the ones named in `systems`), exactly as the user's own click does. Stopping physics puts every body back where it was placed. It stops every project open in that tab. Changes nothing saved. Needs the editor:run right.
| Name | Type | Req | Description |
|---|---|---|---|
| projectId | string | yes | A project open in the user's editor (r3_whoami lists them). |
| systems | array | – | Which of the editor's Run / Stop buttons: physics, code, audio, music. Leave out for all of them. |
| Name | Type | Req | Description |
|---|---|---|---|
| action | string | yes | – |
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| editor | – | yes | – |
| projectId | string | yes | – |
| shown | boolean | – | – |
| systems | array | yes | – |
No examples provided.
r3_editor_undo Undo your own last change to a project ~272
Undo the last change YOU (this agent key or app connection) made to a project: everything one of your tool calls changed is undone together, and the user sees it at once in any R3 tab that has the project open. Use it when your own change was wrong or the user asks you to take it back. It never undoes the user's changes or another agent's: only yours. `count` undoes your last few changes, newest first. If someone else changed the same thing after you, that change of yours is SKIPPED: nothing is changed, the answer says who changed what, and that change of yours is dropped from your history (the next call takes the one before it). Their change stands: setting your old value back over theirs is the user's decision, not this tool's. `redo: true` puts back what you undid (until you make a new change). Needs project:write (and code:write to undo a change to code).
| Name | Type | Req | Description |
|---|---|---|---|
| count | integer | – | How many of your changes to undo (or redo), newest first (default 1, at most 20). |
| projectId | string | yes | The project (r3_project_list). |
| redo | boolean | – | true: redo what you undid instead of undoing. |
| Name | Type | Req | Description |
|---|---|---|---|
| answer | string | yes | This answer in words: what was done, what R3 measured, decided or refused, and what to do next. The same text as the result's text block. Read it: the fields beside it are the data, this is the guida… |
| direction | string | yes | – |
| headSeq | integer | yes | – |
| left | object | yes | – |
| projectId | string | yes | – |
| skippedSteps | array | yes | – |
| steps | array | yes | – |
No examples provided.
How do I install the R3 MCP server?
R3 is a hosted endpoint at https://mcp.r3js.org/mcp, so there is nothing to install locally. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.
remote · mcp.r3js.org
claude mcp add --transport http org-r3js-r3 'https://mcp.r3js.org/mcp'
What is the R3 MCP server?
R3 is an MCP server listed in the public MCP registry as org.r3js/r3. Build 3D web apps: physics, audio, music, Blender import. Live edits, undo, versions, publishing. This page covers its hosted endpoint (https://mcp.r3js.org/mcp).
Is the R3 MCP server safe to use?
R3 scores 94 out of 100 on VerifyMCP. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.
What tools does the R3 MCP server expose?
R3 exposes 80 tools: r3_docs_list, r3_docs_read, r3_whoami, r3_project_create, r3_project_list, and 75 more. Their descriptions and schemas cost roughly 23,934 tokens of context every time the server is loaded.
Does the R3 MCP server require authentication?
Yes. R3 asked us for credentials when we connected, so you will need to authorise it in your MCP client before it can do anything.
Is the R3 MCP server still maintained?
R3 is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.
What did VerifyMCP test on the R3 MCP server?
VerifyMCP ran 14 checks on R3, covering TLS certificate, Authorization, HTTPS, HSTS header and 10 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.