Openrouter Multimodal

Please install the `openrouter-multimodal` MCP server into my current AI client (that's you).

Required prerequisites (do these first if not already done):
- **Install Node.js 18 or newer** — Node.js is required to run the server via npx. The README indicates Node.js 18+. (https://nodejs.org/)
- **Create an OpenRouter API key** — This server requires an OpenRouter API key to function. (https://openrouter.ai/keys)

Canonical MCP server config (stdio transport):
- command: `npx`
- args: ["-y","@stabgan/openrouter-mcp-multimodal"]
- required environment variables:
  - `OPENROUTER_API_KEY`: Your OpenRouter API key. (example: `<your-openrouter-api-key>`)
- optional environment variables:
  - `OPENROUTER_DEFAULT_MODEL`: Default model for chat and analyze tools. (example: `<your-default-model>`)
  - `DEFAULT_MODEL`: Alias for OPENROUTER_DEFAULT_MODEL. (example: `<your-default-model>`)
  - `OPENROUTER_MAX_TOKENS`: Default max_tokens when not set per request. (example: `<your-max-tokens>`)
  - `OPENROUTER_PROVIDER_QUANTIZATIONS`: CSV filter for provider quantizations. (example: `<your-quantizations-csv>`)
  - `OPENROUTER_PROVIDER_IGNORE`: CSV list of provider slugs to exclude. (example: `<your-provider-ignore-csv>`)
  - `OPENROUTER_PROVIDER_SORT`: Provider sort strategy: price, throughput, or latency. (example: `<your-provider-sort>`)
  - `OPENROUTER_PROVIDER_ORDER`: Provider priority order as JSON array or CSV. (example: `<your-provider-order>`)
  - `OPENROUTER_PROVIDER_REQUIRE_PARAMETERS`: Whether providers must require parameters: true or false. (example: `<true-or-false>`)
  - `OPENROUTER_PROVIDER_DATA_COLLECTION`: Provider data collection policy: allow or deny. (example: `<allow-or-deny>`)
  - `OPENROUTER_PROVIDER_ALLOW_FALLBACKS`: Whether provider fallbacks are allowed: true or false. (example: `<true-or-false>`)
  - `OPENROUTER_CACHE_RESPONSES`: Enable response caching server-wide using 1 or true. (example: `<true-or-false>`)
  - `OPENROUTER_INCLUDE_REASONING`: Enable reasoning passthrough server-wide using 1 or true. (example: `<true-or-false>`)
  - `OPENROUTER_MODEL_CACHE_TTL_MS`: Model cache TTL in milliseconds. (example: `<your-cache-ttl-ms>`)
  - `OPENROUTER_IMAGE_MAX_DIMENSION`: Longest edge for image resize in pixels. (example: `<your-image-max-dimension>`)
  - `OPENROUTER_IMAGE_JPEG_QUALITY`: JPEG quality from 1 to 100. (example: `<your-jpeg-quality>`)
  - `OPENROUTER_IMAGE_FETCH_TIMEOUT_MS`: Image URL fetch timeout in milliseconds. (example: `<your-image-fetch-timeout-ms>`)
  - `OPENROUTER_IMAGE_MAX_DOWNLOAD_BYTES`: Maximum image download size in bytes. (example: `<your-image-max-download-bytes>`)
  - `OPENROUTER_IMAGE_MAX_REDIRECTS`: Maximum number of redirects for image URLs. (example: `<your-image-max-redirects>`)
  - `OPENROUTER_IMAGE_MAX_DATA_URL_BYTES`: Maximum image data URL size in bytes. (example: `<your-image-max-data-url-bytes>`)
  - `OPENROUTER_AUDIO_FETCH_TIMEOUT_MS`: Audio URL fetch timeout in milliseconds. (example: `<your-audio-fetch-timeout-ms>`)
  - `OPENROUTER_AUDIO_MAX_DOWNLOAD_BYTES`: Maximum audio download size in bytes. (example: `<your-audio-max-download-bytes>`)
  - `OPENROUTER_AUDIO_MAX_REDIRECTS`: Maximum number of redirects for audio URLs. (example: `<your-audio-max-redirects>`)
  - `OPENROUTER_AUDIO_MAX_DATA_URL_BYTES`: Maximum audio data URL size in bytes. (example: `<your-audio-max-data-url-bytes>`)
  - `OPENROUTER_DEFAULT_VIDEO_MODEL`: Default model for analyze_video. (example: `<your-default-video-model>`)
  - `OPENROUTER_DEFAULT_VIDEO_GEN_MODEL`: Default model for generate_video. (example: `<your-default-video-generation-model>`)
  - `OPENROUTER_VIDEO_FETCH_TIMEOUT_MS`: Video URL fetch timeout in milliseconds. (example: `<your-video-fetch-timeout-ms>`)
  - `OPENROUTER_VIDEO_MAX_DOWNLOAD_BYTES`: Maximum video download size in bytes. (example: `<your-video-max-download-bytes>`)
  - `OPENROUTER_VIDEO_MAX_REDIRECTS`: Maximum number of redirects for video URLs. (example: `<your-video-max-redirects>`)
  - `OPENROUTER_VIDEO_MAX_DATA_URL_BYTES`: Maximum video data URL size in bytes. (example: `<your-video-max-data-url-bytes>`)
  - `OPENROUTER_VIDEO_POLL_INTERVAL_MS`: Polling interval for async video generation jobs in milliseconds. (example: `<your-video-poll-interval-ms>`)
  - `OPENROUTER_VIDEO_MAX_WAIT_MS`: Maximum wait time before returning a resumable video job handle, in milliseconds. (example: `<your-video-max-wait-ms>`)
  - `OPENROUTER_VIDEO_GEN_MAX_BYTES`: Maximum generated video download size in bytes. (example: `<your-video-generation-max-bytes>`)
  - `OPENROUTER_VIDEO_INLINE_MAX_BYTES`: Maximum inline video size in bytes. (example: `<your-video-inline-max-bytes>`)
  - `OPENROUTER_OUTPUT_DIR`: Sandbox root directory for save_path output files. (example: `<your-output-directory>`)
  - `OPENROUTER_ALLOW_UNSAFE_PATHS`: Set to 1 to disable the save_path sandbox. (example: `<1-to-disable-sandbox>`)
  - `OPENROUTER_LOG_LEVEL`: Log level: error, warn, info, or debug. (example: `<your-log-level>`)

Note: Preferred install path is the recommended npx stdio server. The README also documents Docker and global-install alternatives, but npx is the simplest portable option.

Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.