Skip to main content
Glama

EnriVision

EnriVision is a Model Context Protocol (MCP) server over stdio that uploads local media to EnriProxy and returns server-side extraction + model analysis.

This is useful for media types that many MCP clients cannot read reliably (videos, audio, scanned PDFs, HEIC/AVIF, large files), while keeping the MCP server itself lightweight.

What this project is

  • An MCP server process your MCP host launches (OpenCode, Claude Code, Codex, etc.)

  • A thin client for EnriProxy (resumable upload + structured output)

Related MCP server: multimodal-reader-mcp

Requirements

  • Node.js >= 22 (recommended: Node 24 LTS)

  • A reachable EnriProxy server with these endpoints enabled:

    • POST /v1/uploads

    • HEAD /v1/uploads/:id

    • PATCH /v1/uploads/:id

    • POST /v1/vision/analyze

  • An EnriProxy API key (configured on the EnriProxy side)

Install

# Global install
npm install -g @bedolla/enrivision

# Or run without installing
npx -y @bedolla/enrivision@latest --help

Build

npm install
npm run typecheck
npm run build

Usage

1) Configure your MCP host

EnriVision runs as an MCP server over stdio. Your MCP host is responsible for launching the process.

Example: global install

{
  "EnriVision": {
    "type": "stdio",
    "command": "enrivision",
    "args": [],
    "env": {
      "ENRIPROXY_URL": "http://127.0.0.1:8787",
      "ENRIPROXY_API_KEY": "YOUR_ENRIPROXY_API_KEY",
      "ENRIVISION_DEFAULT_LANGUAGE": "es"
    }
  }
}

Example: no install (always uses whatever npm currently tags as latest)

{
  "EnriVision": {
    "type": "stdio",
    "command": "npx",
    "args": ["-y", "@bedolla/enrivision@latest"],
    "env": {
      "ENRIPROXY_URL": "http://127.0.0.1:8787",
      "ENRIPROXY_API_KEY": "YOUR_ENRIPROXY_API_KEY",
      "ENRIVISION_DEFAULT_LANGUAGE": "es"
    }
  }
}
{
  "EnriVision": {
    "type": "stdio",
    "command": "node",
    "args": ["C:\\\\Users\\\\Administrator\\\\Projects\\\\EnriVision\\\\dist\\\\index.js"],
    "env": {
      "ENRIPROXY_URL": "http://127.0.0.1:8787",
      "ENRIPROXY_API_KEY": "YOUR_ENRIPROXY_API_KEY",
      "ENRIVISION_DEFAULT_LANGUAGE": "es"
    }
  }
}

Configuration

EnriVision is configured via environment variables:

  • ENRIPROXY_URL (string, optional, default: http://127.0.0.1:8787)

  • ENRIPROXY_API_KEY (string, required)

  • ENRIVISION_TIMEOUT_MS (string, optional, default: 1800000)

    • Parsed as an integer (milliseconds). Uploads are performed in chunks; this timeout applies per request.

  • ENRIVISION_DEFAULT_LANGUAGE (string, optional)

    • Default language to send when the tool call does not provide language.

MCP tools

EnriVision exposes this MCP tool:

  • analyze_media

General notes:

  • The tool accepts a single JSON object as its input (the MCP arguments).

  • Exactly one of path or paths is required.

  • Paths must be absolute on the machine running the MCP server, or http(s) URLs. URLs are downloaded to a temporary directory on the MCP host (up to 64 MiB each; localhost and private-network destinations are blocked) and deleted after analysis.

  • EnriVision does not accept per-call server_url/api_key overrides (these are configured via env vars).

analyze_media

Inputs:

  • path (string, optional): absolute local file path, or one http(s) URL to download and analyze (up to 64 MiB).

  • paths (string[], optional): absolute local image paths or http(s) image URLs (useful for UI screenshot sets).

  • context (string, optional): high-level hint (examples: ui, diagram, chart, error, code, meeting, tutorial, photo).

  • question (string, optional): what you want to extract/answer.

  • language (string, optional): preferred response language (ISO 639-1; e.g., es, en). If omitted, uses ENRIVISION_DEFAULT_LANGUAGE when set.

  • analysis_mode (string, optional): auto | single | multipass.

  • max_frames (number, optional): single-pass video frames (1..20).

  • transcribe (boolean, optional): enable/disable transcription (videos).

  • transcription_language (string, optional): whisper hint (auto, es, en, ...).

Video targeting:

  • video.clip_start_seconds (number, optional)

  • video.clip_duration_seconds (number, optional)

Multipass tuning (advanced; used only for analysis_mode: multipass):

  • video.segment_seconds (number, optional)

  • video.max_segments (number, optional)

  • video.max_frames_per_segment (number, optional)

  • document.max_pages_total (number, optional)

  • document.pages_per_batch (number, optional)

  • document.max_images_per_batch (number, optional)

  • document.scanned_text_threshold_chars (number, optional)

  • audio.timestamps (boolean, optional)

  • audio.segment_seconds (number, optional)

  • audio.max_segments (number, optional)

  • images.max_images_total (number, optional)

  • images.images_per_batch (number, optional)

  • images.max_dimension (number, optional)

Output:

  • analysis (string): model-produced analysis.

  • media_type (string): detected media type (video, audio, image, document, image_set).

  • extraction (object): safe metadata summary (internal routing details are stripped).

Example arguments object:

{
  "path": "C:\\\\path\\\\to\\\\video.mp4",
  "question": "What are the key steps demonstrated?",
  "analysis_mode": "auto",
  "transcribe": true,
  "language": "es"
}

Many MCP clients include a built-in Read(...) tool that can ingest local files and attach them to the model request. This is convenient, but the set of supported formats is limited and can change across client versions.

If the file you need to analyze is not reliably supported by your client (for example .avif, .heic, .svg, videos, audio, or Office documents), prefer EnriVision MCP so the client can upload bytes and EnriProxy can do extraction reliably.

EnriProxy determines media type using content-type and extension allow-lists.

Videos:

  • .mp4, .mov, .avi, .mkv, .webm, .m4v, .wmv, .flv, .3gp, .3g2, .ts, .mts, .m2ts, .mpeg, .mpg, .gif

Audio:

  • .mp3, .mp1, .mp2, .mpa, .mpga, .wav, .aiff, .aif, .aifc, .caf, .flac, .m4a, .m4b, .m4r, .aac, .ogg, .oga, .wma, .opus, .weba, .mka

Images:

  • .png, .apng, .jpg, .jpeg, .gif, .webp, .avif, .heic, .heif, .tiff, .tif, .bmp, .svg, .ico

Documents:

  • .pdf, .docx, .pptx, .xlsx, .jsonl

Available Tools

1 tool
analyze_mediaA

Sube y analiza un archivo local mediante EnriProxy (extracción del lado servidor + análisis con modelo).

Cuándo usarla:

  • PDFs grandes (muchas páginas) o escaneados donde el Read del cliente puede truncar o perder contenido.

  • Video/audio u otros medios binarios que su cliente no puede leer con Read.

  • Archivos de audio en formatos comunes (mp3, wav, flac, m4a, aac, ogg/oga, opus, wma, weba, mka, aiff/aif/aifc, caf, m4b/m4r, mp1/mp2/mpa/mpga).

  • HEIC/AVIF/TIFF/APNG/SVG/documentos de Office cuando el Read del cliente es poco confiable.

  • Archivos muy grandes que requieren subidas reanudables (hasta 4GB).

  • PDFs/videos grandes: use analysis_mode = 'multipass' para mejor cobertura (auto prefiere multipass para PDFs de más de 20 páginas).

  • Para preguntas de video en un tiempo específico (por ejemplo, "¿qué pasa en 12:34?"), use video.clip_start_seconds y video.clip_duration_seconds.

Reglas:

  • Use path para un archivo, o paths para varias imágenes (capturas de UI/sets de fotos).

  • path/paths aceptan rutas absolutas en la máquina donde corre este servidor MCP (el cliente), o URLs http(s) que se descargan temporalmente en esa misma máquina (hasta 64 MiB; no se permiten hosts locales ni redes privadas).

  • Requiere una API key válida de EnriProxy (env ENRIPROXY_API_KEY, enviada como Authorization: Bearer ...).

  • Prefiera el Read nativo del cliente sólo para texto/PDF/imágenes comunes pequeños y simples cuando funcione; prefiera esta herramienta para PDFs grandes.

  • Responda estrictamente con la salida de la herramienta; si faltan fotogramas/transcripción, dígalo.

  • Video: los fotogramas y la transcripción pertenecen a la MISMA línea de tiempo del video (no son imágenes sin relación).

  • Los GIF/WebP/APNG/SVG animados se convierten en fotogramas clave representativos.

  • Establezca language (por ejemplo, 'es') para coincidir con el idioma del usuario y evitar deriva de idioma.

Depuración de capturas de UI (cuando el material sean capturas de pantalla de aplicaciones):

  • Abra con un veredicto de una línea en lenguaje claro (por ejemplo, 'el formulario de login renderiza correctamente' o 'el header se solapa con la barra lateral').

  • Describa zona por zona (header, barra lateral, contenido principal, modales, notificaciones), no como escena general.

  • Aproxime los colores como valores hex (por ejemplo, #1F6FEB) y nómbrelos; señale colores inesperados o inconsistentes.

  • Cuantifique defectos de layout: desbordes, recortes, solapamientos, desalineaciones, espaciados faltantes, texto cortado; estime magnitudes en píxeles cuando sea posible.

  • Transcriba textualmente etiquetas, botones y cualquier mensaje de error o estado visible.

  • Si la solicitud indica qué se esperaba, compare observado vs esperado de forma explícita.

ParametersJSON Schema
NameRequiredDescriptionDefault
pathNoRuta absoluta a un archivo local en la máquina donde corre el servidor MCP (por ejemplo, C:\\Users\\User\\Downloads\\video.mp4), o una URL http(s) de imagen/video/audio/PDF para descargar y analizar (hasta 64 MiB; hosts locales y redes privadas bloqueados).
audioNoAjuste opcional de multipass para audio (se usa sólo al analizar archivos de audio).
pathsNoRutas absolutas a varios archivos de imagen locales o URLs http(s) (capturas de UI/sets de fotos; cada URL hasta 64 MiB). Cuando se proporcionan, EnriVision sube un único archivo de conjunto para procesamiento por lotes y reducción del lado servidor.
videoNoAjuste opcional de multipass para video. Se usa sólo al analizar videos.
imagesNoAjuste opcional de multipass para conjuntos de imágenes (se usa sólo con `paths`).
regionNoRegión relativa de la IMAGEN original para analizar a resolución nativa (zoom). Coordenadas entre 0 y 1; (0,0) es la esquina superior izquierda. Use las cajas devueltas en 'elements' de un análisis previo de la misma imagen: NUNCA invente coordenadas. Ideal para leer texto pequeño (labels, código) que en la imagen completa comprimida resulta ilegible. Sólo imágenes (path, no paths).
contextNoPista opcional de análisis: ui, diagram, chart, error, code, meeting, tutorial, photo. Déjelo vacío para detección automática.
documentNoAjuste opcional de multipass para documentos (PDF).
languageNoCódigo de idioma preferido de respuesta (ISO 639-1), por ejemplo 'es', 'en'.
questionNoPregunta explícita opcional que responder sobre el archivo.
max_framesNoMáximo opcional de fotogramas para videos (1-20) en modo single-pass. Para tiempos específicos, prefiera video.clip_start_seconds + video.clip_duration_seconds. Para multipass, use video.max_frames_per_segment.
transcribeNoSobreescritura opcional para activar/desactivar la transcripción de audio en videos.
analysis_modeNoSelector opcional de modo de análisis: auto, single o multipass.
transcription_languageNoPista opcional de idioma para la transcripción de audio/video (por ejemplo, 'auto', 'es', 'en').

TDQS

A4.7/5.0
Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries full burden, and it delivers: it discloses auth requirements (ENRIPROXY_API_KEY sent as Bearer), upload/download mechanics (server-side extraction, resumable up to 4GB, URL download up to 64 MiB, private hosts blocked), media handling behavior (animated images converted to keyframes, frames/transcription on the same timeline), and how to report incomplete results ('si faltan fotogramas/transcripción, dígalo').

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is long, but it is well structured with scannable headers ('Cuándo usarla', 'Reglas', 'Depuración de capturas de UI') and bullet points, and it front-loads the core purpose and selection criteria. The UI-debugging block is detailed and somewhat niche, but every section serves a distinct use case, so it earns its place.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 14-parameter tool with no output schema, the description thoroughly covers prerequisites, input types, size limits, auth, multipass modes, and response expectations. It references 'elements' as part of the output, but because no output schema exists, the exact return structure is not fully spelled out—the main remaining gap for an agent trying to use the tool's results.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

The schema already documents all 14 parameters at 100% coverage, so the baseline is 3. The description adds real semantic value by explaining when to use `path` vs `paths`, URL vs local paths, how `video.clip_start_seconds` and `clip_duration_seconds` target timestamped questions, and that `region` coordinates must come from previous 'elements' output rather than being invented.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening sentence states precisely what the tool does: 'Sube y analiza un archivo local mediante EnriProxy (extracción del lado servidor + análisis con modelo).' It identifies the action, resource, and mechanism, and the 'Cuándo usarla' section distinguishes it from the client's native Read by enumerating the specific file types and scenarios for which this tool is the better choice.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines5/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Usage guidance is explicit and actionable: a dedicated 'Cuándo usarla' list gives concrete conditions (large PDFs, audio/video formats, Office docs, up to 4GB), and the rule 'Prefiera el Read nativo del cliente sólo para texto/PDF/imágenes comunes pequeños y simples cuando funcione; prefiera esta herramienta para PDFs grandes' directly names the alternative and when to choose which. Multipass and timestamped video guidance further clarify variant usage.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.

  1. 1 tool updatev0.1.5
    • Changedanalyze_media29 fields changed
      • changedInput schema / properties / analysis_mode / description
        Previous value: -"Optional analysis mode selector: auto, single, or multipass."New value: +"Selector opcional de modo de análisis: auto, single o multipass."
      • changedInput schema / properties / audio / description
        Previous value: -"Optional audio multipass tuning (used only when analyzing audio files)."New value: +"Ajuste opcional de multipass para audio (se usa sólo al analizar archivos de audio)."
      • changedInput schema / properties / audio / properties / max_segments / description
        Previous value: -"Maximum number of audio segments to analyze."New value: +"Número máximo de segmentos de audio a analizar."
      • changedInput schema / properties / audio / properties / segment_seconds / description
        Previous value: -"Segment duration in seconds for audio multipass."New value: +"Duración del segmento en segundos para multipass de audio."
      • changedInput schema / properties / audio / properties / timestamps / description
        Previous value: -"Whether to include timestamped segments in audio extraction."New value: +"Si incluir segmentos con marca de tiempo en la extracción de audio."
      • changedInput schema / properties / context / description
        Previous value: -"Optional analysis hint: ui, diagram, chart, error, code, meeting, tutorial, photo. Leave empty for auto-detection."New value: +"Pista opcional de análisis: ui, diagram, chart, error, code, meeting, tutorial, photo. Déjelo vacío para detección automática."
      • changedInput schema / properties / document / description
        Previous value: -"Optional document multipass tuning (PDF)."New value: +"Ajuste opcional de multipass para documentos (PDF)."
      • changedInput schema / properties / document / properties / max_images_per_batch / description
        Previous value: -"Maximum rendered pages (images) per batch."New value: +"Máximo de páginas renderizadas (imágenes) por lote."
      • changedInput schema / properties / document / properties / max_pages_total / description
        Previous value: -"Maximum number of pages to analyze in total."New value: +"Número máximo de páginas a analizar en total."
      • changedInput schema / properties / document / properties / pages_per_batch / description
        Previous value: -"Pages per batch for multipass map calls."New value: +"Páginas por lote para las llamadas map de multipass."
      • changedInput schema / properties / document / properties / scanned_text_threshold_chars / description
        Previous value: -"Minimum extracted text length to treat a page as textual."New value: +"Longitud mínima de texto extraído para tratar una página como textual."
      • changedInput schema / properties / images / description
        Previous value: -"Optional image-set multipass tuning (used only with `paths`)."New value: +"Ajuste opcional de multipass para conjuntos de imágenes (se usa sólo con `paths`)."
      • changedInput schema / properties / images / properties / images_per_batch / description
        Previous value: -"Images per batch for multipass map calls."New value: +"Imágenes por lote para las llamadas map de multipass."
      • changedInput schema / properties / images / properties / max_dimension / description
        Previous value: -"Maximum dimension for images (width/height)."New value: +"Dimensión máxima para las imágenes (ancho/alto)."
      • changedInput schema / properties / images / properties / max_images_total / description
        Previous value: -"Maximum number of images to analyze in total."New value: +"Número máximo de imágenes a analizar en total."
      • changedInput schema / properties / language / description
        Previous value: -"Preferred response language code (ISO 639-1), e.g. 'es', 'en'."New value: +"Código de idioma preferido de respuesta (ISO 639-1), por ejemplo 'es', 'en'."
      • changedInput schema / properties / max_frames / description
        Previous value: -"Optional max frames for videos (1-20) in single-pass mode. For targeted timestamps, prefer video.clip_start_seconds + video.clip_duration_seconds. For multipass, use video.max_frames_per_segment."New value: +"Máximo opcional de fotogramas para videos (1-20) en modo single-pass. Para tiempos específicos, prefiera video.clip_start_seconds + video.clip_duration_seconds. Para multipass, use video.max_frames_per_segment."
      • changedInput schema / properties / path / description
        Previous value: -"Absolute path to a local file on the machine running the MCP server (e.g., C:\\\\Users\\\\User\\\\Downloads\\\\video.mp4)."New value: +"Ruta absoluta a un archivo local en la máquina donde corre el servidor MCP (por ejemplo, C:\\\\Users\\\\User\\\\Downloads\\\\video.mp4), o una URL http(s) de imagen/video/audio/PDF para descargar y analizar (hasta 64 MiB; hosts locales y redes privadas bloqueados)."
      • changedInput schema / properties / paths / description
        Previous value: -"Absolute paths to multiple local image files (UI screenshots/photo sets). When provided, EnriVision uploads a single media-set archive for server-side batching + reduce."New value: +"Rutas absolutas a varios archivos de imagen locales o URLs http(s) (capturas de UI/sets de fotos; cada URL hasta 64 MiB). Cuando se proporcionan, EnriVision sube un único archivo de conjunto para procesamiento por lotes y reducción del lado servidor."
      • changedInput schema / properties / question / description
        Previous value: -"Optional explicit question to answer about the file."New value: +"Pregunta explícita opcional que responder sobre el archivo."
      • addedInput schema / properties / region
        Added value: +{
        +  "description": "Región relativa de la IMAGEN original para analizar a resolución nativa (zoom). Coordenadas entre 0 y 1; (0,0) es la esquina superior izquierda. Use las cajas devueltas en 'elements' de un análisis previo de la misma imagen: NUNCA invente coordenadas. Ideal para leer texto pequeño (labels, código) que en la imagen completa comprimida resulta ilegible. Sólo imágenes (path, no paths).",
        +  "properties": {
        +    "height": {
        +      "description": "Alto relativo (1 = alto completo).",
        +      "type": "number"
        +    },
        +    "width": {
        +      "description": "Ancho relativo (1 = ancho completo).",
        +      "type": "number"
        +    },
        +    "x": {
        +      "description": "Coordenada horizontal relativa de la esquina superior izquierda (0 = borde izquierdo).",
        +      "type": "number"
        +    },
        +    "y": {
        +      "description": "Coordenada vertical relativa de la esquina superior izquierda (0 = borde superior).",
        +      "type": "number"
        +    }
        +  },
        +  "required": [
        +    "x",
        +    "y",
        +    "width",
        +    "height"
        +  ],
        +  "type": "object"
        +}
      • changedInput schema / properties / transcribe / description
        Previous value: -"Optional override to enable/disable audio transcription for videos."New value: +"Sobreescritura opcional para activar/desactivar la transcripción de audio en videos."
      • changedInput schema / properties / transcription_language / description
        Previous value: -"Optional Whisper language hint for audio/video transcription (e.g., 'auto', 'es', 'en')."New value: +"Pista opcional de idioma para la transcripción de audio/video (por ejemplo, 'auto', 'es', 'en')."
      • changedInput schema / properties / video / description
        Previous value: -"Optional video multipass tuning. Used only when analyzing videos."New value: +"Ajuste opcional de multipass para video. Se usa sólo al analizar videos."
      • changedInput schema / properties / video / properties / clip_duration_seconds / description
        Previous value: -"Optional clip duration in seconds for time-targeted video analysis."New value: +"Duración opcional del clip en segundos para análisis de video dirigido a un tiempo."
      • changedInput schema / properties / video / properties / clip_start_seconds / description
        Previous value: -"Optional clip start offset in seconds for time-targeted video analysis."New value: +"Offset opcional de inicio del clip en segundos para análisis de video dirigido a un tiempo."
      • changedInput schema / properties / video / properties / max_frames_per_segment / description
        Previous value: -"Maximum frames to extract per segment."New value: +"Máximo de fotogramas a extraer por segmento."
      • changedInput schema / properties / video / properties / max_segments / description
        Previous value: -"Maximum number of segments to analyze."New value: +"Número máximo de segmentos a analizar."
      • changedInput schema / properties / video / properties / segment_seconds / description
        Previous value: -"Segment duration in seconds."New value: +"Duración del segmento en segundos."
  2. 1 tool updatev0.1.1
    • Changedanalyze_media1 field changed
      • addedInput schema / properties / analysis_mode / enum
        Added value: +[
        +  "auto",
        +  "single",
        +  "multipass"
        +]
  3. 1 tool updatev0.1.0
    • First observedanalyze_media

TDQS

A4.6/5.0
Disambiguation5/5

With only one tool, there is no possibility of confusing it with other tools. The tool's purpose is clearly defined and self-contained.

Naming Consistency5/5

The single tool name 'analyze_media' follows a clear verb_noun pattern. There are no other names to create inconsistency.

Tool Count3/5

The server has only one tool, which feels thin for a media analysis service. However, the tool is extremely comprehensive, handling many file types and modes, making the minimal count reasonable but still borderline.

Completeness5/5

The tool covers a complete analysis workflow for the server's stated purpose: upload, extraction, analysis, and support for many media types. No obvious gaps are present.

Maintenance

ActivityMaintained
ResponsivenessNo issues

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Connectors

Related MCP Servers

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/Bedolla/EnriVision'

If you have feedback or need assistance with the MCP directory API, please join our Discord server