EasyAR MCP Server
Enables AI coding tools like Codex and Claude to assist with EasyAR Unity workflows, including project inspection, sample preparation, and C# script generation.
README
<p align="center"> <img src="assets/easyar-icon.png" alt="EasyAR logo" width="96" height="96"> </p>
mcp-easyar
mcp-easyar helps registered EasyAR users connect AI coding tools such as Codex and Claude to authorized EasyAR Unity workflows.
This MVP focuses on local Unity project assistance. The current sample run-through scope is intentionally limited to Image Tracking and Cloud Recognition; Hello AR, Surface Tracking, and other samples are cataloged for later expansion.
- inspect Unity project structure and EasyAR-related files
- report server status, capability summary, resources, and recommended first calls
- diagnose whether a Unity project is ready to run a selected EasyAR sample
- generate official EasyAR Unity Plugin and focused sample import checklists
- write import checklists into Unity projects as Markdown artifacts
- generate an ordered focused sample run sequence for Codex or Claude
- write focused run sequences into Unity projects as Markdown artifacts
- index focused sample handoff artifacts with recommended reading order
- generate a focused sample run report across readiness, config, and script review
- write focused run reports into Unity projects as Markdown artifacts
- audit focused sample scenes, Build Settings hints, EasyAR import signals, and sample-specific blockers
- recognize Image Tracking scenes by scene content markers as well as official sample naming hints
- report EasyAR PackageCache
Samples~candidates when focused sample scenes have not been imported intoAssets/Samples - generate Unity Package Manager sample import guides for moving focused samples from
Samples~intoAssets/Samples - generate focused support bundles that summarize project state, scene audit, run report, and latest Unity log diagnostics
- generate real-device validation checklists for focused Image Tracking and Cloud Recognition tests
- write device validation checklists into Unity projects as Markdown artifacts
- record focused sample run results for compile, build, and device validation handoff
- generate redacted GitHub issue reports for focused sample failures
- plan scoped Unity C# changes before Codex or Claude writes sample scripts
- summarize Unity C# code changes after script edits and before Unity compilation
- list common EasyAR Unity sample categories with focused/deferred status
- generate an Editor runner script for opening sample scenes
- generate an Editor helper for Android/iOS player settings and camera permissions
- generate an Editor helper for adding sample scenes to Unity Build Settings
- generate an Editor helper for Unity-side focused sample validation
- generate an Editor helper for Android/iOS/Standalone player builds
- generate focused runbooks and support directories for Image Tracking and Cloud Recognition
- generate a single focused preflight gate across account, local config, Unity path, imports, scene readiness, and script review
- prepare a Unity project with EasyAR local config templates and secret ignore rules
- validate local EasyAR config files without exposing secrets
- write local EasyAR Unity config from environment-backed secrets without returning secret values
- generate focused Unity programming context before editing C# scripts
- generate Unity MonoBehaviour script templates for common EasyAR workflows
- write C# scripts safely inside a Unity project
- review Unity C# scripts for common EasyAR and AR sample risks
- run a Unity batch compile/import check after code changes
- run a Unity static editor method in batch mode
- inspect Unity executable configuration before batch mode runs
- write Unity executable setup reports with
EASYAR_UNITY_PATHguidance and compile dry-run commands - analyze Unity Editor/build logs for common EasyAR issues plus Image Tracking and Cloud Recognition specifics
- find and analyze the latest Unity log after a failed batch run
- expose EasyAR workflow guidance as MCP resources
- expose MCP prompts for Image Tracking, Cloud Recognition, and Unity programming assistance
- guide unregistered users through official EasyAR registration, development center login, license creation, and Cloud Recognition credential setup without collecting passwords
- generate account material checklists that classify EasyAR fields by source, storage location, and share policy
- check Codex/Claude/generic MCP client setup before users paste config
- write client setup reports for handoff and troubleshooting
- recommend the next focused workflow step from current project evidence
- write workflow state handoff artifacts for Codex/Claude continuation
- check EasyAR account environment configuration without exposing secrets
- call configured official EasyAR account/license validation endpoints without exposing secrets
- run focused official access checks across account, license, downloads, and Cloud Recognition endpoints
- write official access reports into Unity projects as Markdown artifacts
- generate MCP client configuration snippets for Codex and Claude
- generate and write deployment readiness reports for GitHub release, official endpoint, Unity path, and security checks
- generate release/install manifests for users and AI tools
- generate first-run onboarding reports across client setup, official access, release metadata, and focused workflow state
The server is intentionally built for official, authorized use. Production deployments should connect EASYAR_API_BASE_URL and EASYAR_API_TOKEN to EasyAR account/license APIs before serving private SDK downloads or account-scoped content.
This project must not be used to bypass EasyAR login, license checks, download authorization, enterprise gates, rate limits, or any other access control.
Official references used by this MVP:
- EasyAR sample apps: https://www.easyar.cn/doc/en/develop/samples.html
- EasyAR download page: https://www.easyar.com/view/download.html
- EasyAR download history: https://www.easyar.com/view/downloadHistory.html
Install
npm install
npm run build
For package-based clients after publishing, use npm install -g mcp-easyar with entrypointMode=package-bin, or entrypointMode=npx for npx -y mcp-easyar.
After installing the package, verify the MCP entrypoint without exposing account secrets:
easyar-mcp-check
For a local build, use npm run install:check.
Run
npm start
Optional environment variables:
EASYAR_API_BASE_URL=https://www.easyar.cn
EASYAR_API_TOKEN=your_registered_user_token
EASYAR_ACCOUNT_STATUS_ENDPOINT=https://www.easyar.cn/path/to/official/account/status
EASYAR_LICENSE_VALIDATE_ENDPOINT=https://www.easyar.cn/path/to/official/license/validate
EASYAR_DOWNLOADS_ENDPOINT=https://www.easyar.cn/path/to/official/downloads
EASYAR_CLOUD_CREDENTIALS_ENDPOINT=https://www.easyar.cn/path/to/official/cloud-recognition/credentials
EASYAR_UNITY_PATH=/Applications/Unity/Hub/Editor/2022.3.62f1/Unity.app/Contents/MacOS/Unity
EASYAR_UNITY_CANDIDATE_DIRS=/Applications/Unity/Hub/Editor
Use .env.example as a non-secret template. Keep real .env files, tokens, license keys, and Cloud Recognition credentials local.
EASYAR_ACCOUNT_STATUS_ENDPOINT, EASYAR_LICENSE_VALIDATE_ENDPOINT, EASYAR_DOWNLOADS_ENDPOINT, and EASYAR_CLOUD_CREDENTIALS_ENDPOINT are intentionally configurable. Connect them to authorized EasyAR account APIs in production; the open-source default does not guess or bypass private EasyAR endpoints.
See docs/OFFICIAL_API_CONTRACT.md or the machine-readable docs/openapi/easyar-mcp-account-api.openapi.json for the required official account API contract, request fields, response policy, and secret-handling rules. MCP clients can also call easyar_generate_official_api_contract.
Claude Desktop
{
"mcpServers": {
"easyar": {
"command": "node",
"args": ["/absolute/path/to/mcp-easyar/dist/index.js"],
"env": {
"EASYAR_API_BASE_URL": "https://www.easyar.cn",
"EASYAR_API_TOKEN": "your_registered_user_token"
}
}
}
}
Codex
Use the same stdio command:
node /absolute/path/to/mcp-easyar/dist/index.js
You can also ask the MCP server for a config snippet with easyar_generate_client_config. The default entrypointMode=local-dist uses node /absolute/path/to/dist/index.js; entrypointMode=package-bin uses the installed easyar-mcp bin; entrypointMode=npx uses npx -y mcp-easyar.
Local Development
npm install
npm run dev
Run checks:
npm run build
npm run typecheck
npm test
npm run bin:smoke
npm run install:check
npm run package:smoke
npm run pack:check
Quickstart
See docs/quickstart.md for the recommended Codex/Claude to Unity workflow.
See docs/client-setup.md for Codex, Claude Desktop, local-dist, package-bin, and npx setup profiles.
See docs/troubleshooting.md for focused Image Tracking and Cloud Recognition diagnostics.
Release Manifest
See docs/RELEASE_MANIFEST.md for install commands, MCP entrypoints, focused scope, environment variable names, and release checks.
Release Checks
Before publishing a release or npm package:
npm run typecheck
npm test
npm run bin:smoke
npm run install:check
npm run package:smoke
npm run pack:check
npm run security:check
npm run release:check
release:check runs the package/repository verification commands and then calls easyar_production_validation. It exits successfully while production evidence is still incomplete, but prints the blockers. Set EASYAR_RELEASE_REQUIRE_PRODUCTION_READY=1 before a real release tag or npm publish to make incomplete production readiness fail the command.
Npm publishing should use the manual GitHub Actions Release workflow. It runs the strict production gate first, then publishes with npm provenance from the protected npm-publish environment.
See CHANGELOG.md for version history.
First-Time EasyAR Users
mcp-easyar treats account setup as a guided browser handoff. If the user has not registered yet, start with easyar_first_run_guide accountStage=not-registered sampleId=cloud-recognition; the guide chooses the first safe MCP call, sends the user to the official EasyAR registration/login pages, and explains what to do after they return to Codex or Claude. The MCP server never asks for website passwords, verification codes, raw API tokens, license keys, API keys, appKey, or appSecret in chat.
After registration/login, the user creates or locates the EasyAR Sense license and, for Cloud Recognition, the CRS/Cloud Recognition AppId plus API KEY. Sense 4.1+ uses easyar.cloudRecognition.appId + easyar.cloudRecognition.apiKey; legacy appKey/appSecret fields remain accepted for compatibility. Those values are filled locally in ProjectSettings/EasyAR/easyar.local.json; easyar_validate_local_config reports only whether required fields are present and non-placeholder.
For Unity-project handoff, call easyar_write_first_run_guide projectPath=/path/to/UnityProject accountStage=not-registered sampleId=cloud-recognition, then easyar_write_local_config_handoff projectPath=/path/to/UnityProject accountStage=not-registered sampleId=cloud-recognition. These write Assets/EasyARGenerated/FIRST_RUN.md and Assets/EasyARGenerated/LOCAL_CONFIG_HANDOFF.md, combining the official browser registration/login path, focused sample scope, account materials checklist, manual easyar.local.json steps, environment-backed writer command, and validation chain.
For brand-new users, the intended conversation is:
- MCP asks one lightweight question: "Do you already have an EasyAR account?"
- If the answer is no, MCP uses
accountStage=not-registeredand gives browser-only actions: openhttps://www.easyar.cn/, register from the official login/register entry, activate the email if required, then enter the development center. - The user returns with only a stage update, such as
registered-not-logged-inorlogged-in; they do not paste credentials into chat. - MCP guides license creation, package/bundle identifier matching, official sample/download access, and Cloud Recognition credential setup when the selected sample needs it.
- Secret values are written locally or supplied from environment variables; MCP validates presence and placeholders, then moves to focused preflight.
The first MCP screen is intentionally account-stage driven:
easyar_server_statusshowspreflightFirst=trueand suggestseasyar_first_run_guide accountStage=not-registeredas a valid starting point.easyar_write_first_run_guidecreatesFIRST_RUN.md, the first-screen route for new users, focused scope, top next call, artifact reading order, and Unity-automation gate.easyar_write_account_onboardingcreatesACCOUNT_ONBOARDING.mdso another AI tool or human operator can resume the same browser handoff.easyar_write_account_materialslists which official materials are needed, where to find them, where to store them locally, and whether they are safe to share.easyar_write_local_config_handoffcreatesLOCAL_CONFIG_HANDOFF.md, the safest bridge from EasyAR website registration/login to local secret storage.easyar_write_local_config_formcreatesLOCAL_CONFIG_FORM.md, a field-by-field local JSON fill form with placeholders, source pages, env alternatives, and validation calls.easyar_prepare_unity_projectcreates the localeasyar.local.json.examplewith field sources, env-variable alternatives, and a never-share list.easyar_validate_local_configchecks only presence and placeholders.easyar_write_focused_preflightcreatesPREFLIGHT.md, the gate before Unity batch or device validation.easyar_write_focused_handoff_packwrites the safe handoff pack for Image Tracking, Cloud Recognition, or both focused samples so another AI tool can continue fromHANDOFF_PACK.mdandARTIFACT_INDEX.md.- Before recording device evidence,
easyar_write_device_run_result_formcreatesDEVICE_RUN_RESULT_FORM.md, a fillable real-device evidence form plus safeeasyar_write_run_resultargument templates. - After compile/build/device attempts,
easyar_write_completion_reportcreatesCOMPLETION_REPORT.md;runThroughComplete=trueonly when preflight, device validation, a passedRUN_RESULT.mdwith recorded real-device validation evidence, and latest log evidence all support a passed focused run-through.
Tools
easyar_list_samples: list supported sample categories.easyar_server_status: return server version, capability summary, resources, authorization state, and recommended first calls.easyar_official_info: return official EasyAR links and currently documented package versions captured by this MCP server.easyar_auth_status: check EasyAR API environment configuration without returning secrets.easyar_check_account: call a configured official EasyAR account-status endpoint withEASYAR_API_TOKEN.easyar_validate_license: call a configured official EasyAR license-validation endpoint using local config or explicit input without returning secrets.easyar_discover_downloads: call a configured official downloads endpoint for account-scoped SDK/sample package discovery without returning tokens.easyar_discover_cloud_credentials: call a configured official Cloud Recognition endpoint for account-scoped credential metadata without returning secrets.easyar_generate_official_api_contract: generate the official EasyAR account API contract needed for production endpoint integration.easyar_write_official_api_contract: write the official API contract todocs/OFFICIAL_API_CONTRACT.mdor a selected workspace path.easyar_generate_official_openapi_contract: return the machine-readable official OpenAPI contract for gateway import, backend stubs, or client generation.easyar_write_official_openapi_contract: write the official OpenAPI contract JSON todocs/openapi/easyar-mcp-account-api.openapi.jsonor a selected workspace path.easyar_official_api_handoff: generate a backend/operations handoff for wiring real EasyAR official account APIs into mcp-easyar.easyar_write_official_api_handoff: write the official API rollout handoff todocs/OFFICIAL_API_HANDOFF.mdor a selected workspace path.easyar_check_official_access: run a focused official account, license, downloads, and sample-specific access check without exposing secrets.easyar_write_official_access_report: write the official access check toAssets/EasyARGenerated/<sampleId>/OFFICIAL_ACCESS.md.easyar_generate_client_config: generate Codex, Claude Desktop, or generic MCP stdio configuration.easyar_check_client_setup: validate Node.js, package/bin metadata, built entrypoint, server path, and generated MCP client config.easyar_write_client_setup: write client setup status, config snippet, acceptance checklist, first smoke calls, and troubleshooting steps toEasyARGenerated/CLIENT_SETUP.mdorAssets/EasyARGenerated/CLIENT_SETUP.md.easyar_deployment_readiness: generate a deployment readiness report across package metadata, GitHub release files, official endpoint environment, Unity path, focused scope, and security checks.easyar_write_deployment_readiness: write the deployment readiness report as Markdown inside a project or workspace directory.easyar_production_validation: generate the final production evidence matrix across release manifest, deployment readiness, official EasyAR endpoints, verification commands, and focused sample completion.easyar_write_production_validation: write the production validation evidence matrix toPRODUCTION_VALIDATION.md; it marks the deployment incomplete unless real evidence is present.easyar_release_manifest: generate a consumer-facing install/release manifest with package metadata, entrypoints, focused scope, env names, and verification commands.easyar_write_release_manifest: write the release manifest todocs/RELEASE_MANIFEST.mdor a selected path.easyar_first_run_guide: generate the first-screen guide for new MCP users across account stage, official browser route, focused scope, local config, and the first safe MCP calls.easyar_write_first_run_guide: writeFIRST_RUN.mdtoAssets/EasyARGenerated/FIRST_RUN.mdorEasyARGenerated/FIRST_RUN.md.easyar_account_onboarding: guide new or existing EasyAR users through official registration/login, license setup, Cloud Recognition credentials, and local MCP setup without collecting passwords or secrets.easyar_write_account_onboarding: write the account onboarding guide toAssets/EasyARGenerated/ACCOUNT_ONBOARDING.mdorEasyARGenerated/ACCOUNT_ONBOARDING.md.easyar_account_materials: generate a field-by-field checklist of EasyAR account materials, source locations, local storage targets, and share policies.easyar_write_account_materials: write the account materials checklist toAssets/EasyARGenerated/ACCOUNT_MATERIALS.mdorEasyARGenerated/ACCOUNT_MATERIALS.md.easyar_portal_evidence: generate a safe EasyAR portal evidence report from non-secret browser observations such as app record id, service flags, license presence, and Cloud Recognition library status.easyar_write_portal_evidence: write the safe portal evidence report toAssets/EasyARGenerated/PORTAL_EVIDENCE.mdorEasyARGenerated/PORTAL_EVIDENCE.mdwithout storing API KEY/API Secret/license values.easyar_onboarding_report: generate a first-run project onboarding report across client setup, official access, release manifest, and workflow state.easyar_write_onboarding_report: write the onboarding report toAssets/EasyARGenerated/<sampleId>/ONBOARDING.md.easyar_generate_project_handoff: generate a project-level continuation dashboard across client setup, account materials, local config, Unity environment, focused sample state, and next calls.easyar_write_project_handoff: write the project continuation dashboard toAssets/EasyARGenerated/PROJECT_HANDOFF.md.easyar_remaining_work_report: generate an evidence-weighted remaining-work report across repository release surface, official EasyAR APIs, Unity config, and focused sample run-through evidence.easyar_write_remaining_work_report: write the remaining-work gap report toAssets/EasyARGenerated/REMAINING_WORK.mdorEasyARGenerated/REMAINING_WORK.md.easyar_generate_focused_preflight: generate a single focused preflight gate across account materials, local config, Unity executable setup, sample imports, scene readiness, and script review.easyar_write_focused_preflight: write the focused preflight gate toAssets/EasyARGenerated/<sampleId>/PREFLIGHT.md.easyar_next_workflow_step: inspect current focused project state and recommend the next MCP/Unity action.easyar_write_workflow_state: write current workflow phase, blockers, and next call toAssets/EasyARGenerated/<sampleId>/WORKFLOW_STATE.md.easyar_generate_import_checklist: generate an official EasyAR Unity Plugin and focused sample import checklist for Image Tracking or Cloud Recognition.easyar_write_import_checklist: write the official import checklist toAssets/EasyARGenerated/<sampleId>/IMPORT_CHECKLIST.md.easyar_generate_sample_import_guide: generate Unity Package Manager steps, expected import locations, and post-import verification calls for importing the focused official sample from PackageCacheSamples~intoAssets/Samples.easyar_write_sample_import_guide: write the focused sample import guide toAssets/EasyARGenerated/<sampleId>/SAMPLE_IMPORT_GUIDE.md.easyar_import_sample_from_package_cache: copy a focused sample already present in local Unity PackageCacheSamples~intoAssets/Sampleswithout downloading packages or exposing secrets.easyar_generate_run_sequence: generate an ordered MCP/Unity batch sequence for Image Tracking or Cloud Recognition.easyar_write_run_sequence: write the focused run sequence toAssets/EasyARGenerated/<sampleId>/RUN_SEQUENCE.md.easyar_generate_artifact_index: generate an index of focused handoff artifacts, including account onboarding/materials, and recommended reading order.easyar_write_artifact_index: write the focused artifact index toAssets/EasyARGenerated/<sampleId>/ARTIFACT_INDEX.md.easyar_generate_focused_handoff_pack: preview the safe focused sample handoff pack that can be written for Image Tracking, Cloud Recognition, or both.easyar_write_focused_handoff_pack: write the focused handoff pack (HANDOFF_PACK.md,ARTIFACT_INDEX.md, diagnostics, forms, and project dashboards) without writing secrets or fake run results.easyar_generate_run_report: summarize focused sample readiness, local config validation, script review, and next phase.easyar_write_run_report: write the focused sample run report toAssets/EasyARGenerated/<sampleId>/RUN_REPORT.md.easyar_audit_sample_scene: audit scene candidates, Build Settings hints, EasyAR import signals, and focused sample blockers.easyar_write_scene_audit: write the focused scene audit toAssets/EasyARGenerated/<sampleId>/SCENE_AUDIT.md.easyar_generate_support_bundle: generate a focused support bundle summary across run sequence, run report, scene audit, and latest Unity log diagnostics.easyar_write_support_bundle: write the focused support bundle toAssets/EasyARGenerated/<sampleId>/SUPPORT_BUNDLE.md.easyar_generate_device_validation_checklist: generate focused real-device validation steps, pass criteria, blockers, and evidence prompts.easyar_write_device_validation_checklist: write the real-device validation checklist toAssets/EasyARGenerated/<sampleId>/DEVICE_VALIDATION.md.easyar_generate_device_run_result_form: generate a fillable real-device run result form plus safeeasyar_write_run_resulttemplates.easyar_write_device_run_result_form: write the real-device run result form toAssets/EasyARGenerated/<sampleId>/DEVICE_RUN_RESULT_FORM.md.easyar_generate_run_result: generate a focused sample run result summary for compile, build, or device validation attempts.easyar_write_run_result: write the focused run result toAssets/EasyARGenerated/<sampleId>/RUN_RESULT.md.easyar_generate_completion_report: summarize final focused sample completion status across preflight, device validation,RUN_RESULT.md, and latest Unity log evidence.easyar_write_completion_report: write the final focused completion report toAssets/EasyARGenerated/<sampleId>/COMPLETION_REPORT.md.easyar_generate_focused_scope_status: aggregate Image Tracking and Cloud Recognition completion states into one focused-scope status.easyar_write_focused_scope_status: write the focused-scope status toAssets/EasyARGenerated/FOCUSED_SCOPE_STATUS.md.easyar_generate_issue_report: generate a redacted GitHub issue body for focused Image Tracking or Cloud Recognition failures.easyar_write_issue_report: write the redacted issue report toAssets/EasyARGenerated/<sampleId>/ISSUE_REPORT.md.easyar_inspect_unity_project: inspect a Unity project for package, asset, scene, and EasyAR signals.easyar_check_sample_readiness: report missing local requirements before running a selected sample.easyar_validate_local_config: validateProjectSettings/EasyAR/easyar.local.jsonwithout returning secrets.easyar_generate_local_config_form: generate a fillableeasyar.local.jsonform with field sources, placeholders, environment alternatives, and validation calls.easyar_write_local_config_form: write the local config form toAssets/EasyARGenerated/LOCAL_CONFIG_FORM.mdwithout writing secret values.easyar_write_local_config_from_env: writeProjectSettings/EasyAR/easyar.local.jsonfrom local environment variables such asEASYAR_API_TOKEN,EASYAR_LICENSE_KEY, and Cloud Recognition env vars without returning secret values.easyar_local_config_handoff: generate a first-run handoff from EasyAR website registration/login to safe localeasyar.local.jsonsetup and validation.easyar_write_local_config_handoff: write the first-run local config handoff toAssets/EasyARGenerated/LOCAL_CONFIG_HANDOFF.md.easyar_generate_sample_plan: create a practical setup/run plan for a sample.easyar_prepare_unity_project: create a sample runner, focused sample runbook, EasyAR local config template, and secret ignore rules in a Unity project.easyar_create_sample_validation_helper: createAssets/Editor/EasyARSampleValidationHelper.csfor Unity-side focused sample validation, excluding generated MCP helper files from official EasyAR import signals.easyar_create_local_config_bridge: create Editor/runtime scripts that export a minimized EasyAR runtime config into ignoredStreamingAssetsfor device builds and read it at runtime without logging secret values or copying account tokens/API secrets.easyar_create_mobile_settings_helper: createAssets/Editor/EasyARMobileSettingsHelper.csfor Android/iOS player settings and camera permission setup.easyar_create_build_settings_helper: createAssets/Editor/EasyARBuildSettingsHelper.csfor Build Settings automation.easyar_create_device_build_helper: createAssets/Editor/EasyARDeviceBuildHelper.csfor Unity player builds.easyar_create_sample_runner: createAssets/Editor/EasyARSampleRunner.cs.easyar_generate_config_integration_audit: audit how local EasyAR config can be wired into Unity scripts, scenes, prefabs, and assets without exposing secret values.easyar_write_config_integration_audit: write the local-config integration audit toAssets/EasyARGenerated/<sampleId>/CONFIG_INTEGRATION.md.easyar_generate_programming_context: generate a focused Unity C# script inventory, generated helper summary, static review, and recommended programming workflow before editing scripts.easyar_write_programming_context: write the focused programming context toAssets/EasyARGenerated/<sampleId>/PROGRAMMING_CONTEXT.md.easyar_generate_code_plan: generate a focused Unity C# implementation plan before editing sample code, including verification calls for script review, Unity compile, support bundle, and run result handoff.easyar_write_code_plan: write the focused C# implementation plan and verification calls toAssets/EasyARGenerated/<sampleId>/CODE_PLAN.md.easyar_create_mono_behaviour: create a Unity C# MonoBehaviour template for EasyAR sample development.easyar_write_csharp_file: create or replace a.csfile inside a Unity project.easyar_generate_code_change_summary: generate a focused C# change summary after script edits.easyar_write_code_change_summary: write the focused C# change summary toAssets/EasyARGenerated/<sampleId>/CODE_CHANGE.md.easyar_review_csharp_scripts: review Unity C# scripts for common EasyAR secret, lifecycle, input, and performance risks.easyar_unity_environment: inspectEASYAR_UNITY_PATHand common Unity install locations without launching Unity.easyar_write_unity_environment_report: write Unity executable discovery,EASYAR_UNITY_PATH, and batch compile dry-run guidance toAssets/EasyARGenerated/UNITY_ENVIRONMENT.md.easyar_run_unity_compile_check: open Unity in batch mode to force script import/compilation, analyze the generated log, and return a suggestedeasyar_write_run_resulthandoff whensampleIdis provided.easyar_run_unity_method: execute a Unity static editor method in batch mode, optionally writing a project-local Unity log withlogPath; passsampleIdto receive focused log diagnostics and a suggestedRUN_RESULT.mdupdate.easyar_android_device_status: inspect adb availability and connected Android devices before focused real-device validation.easyar_android_install_apk: install a focused Android APK with adb and return a safeRUN_RESULT.mdhandoff step.easyar_android_start_app: launch the focused Android app by package name with adb monkey.easyar_android_collect_logcat: collect a filtered, redacted adb logcat snapshot into the Unity project for device evidence.easyar_analyze_unity_log: analyze Unity Editor/build logs and suggest next debugging actions, with optional focused sample diagnostics.easyar_analyze_latest_unity_log: locate the latest Unity Editor/project log and analyze its tail.
Resources
easyar://samples/catalogeasyar://official/infoeasyar://official/api-contracteasyar://official/openapieasyar://unity/checklisteasyar://workflow/quickstarteasyar://workflow/focused-scopeeasyar://workflow/programming
Prompts
easyar-run-image-trackingeasyar-run-cloud-recognitioneasyar-validate-official-endpointseasyar-close-focused-scopeeasyar-unity-programming-assistant
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.