> ## Documentation Index > Fetch the complete documentation index at: https://docs.shareofmodel.ai/llms.txt > Use this file to discover all available pages before exploring further. # Create asset evaluation > Create a new asset evaluation over assets already in the workspace's library. The assets are referenced by `assets_ids` and must all belong to the workspace in the path; referencing an asset from another workspace is rejected with `400`. **Content types.** An evaluation may span more than one content type — for example an image and a video in the same evaluation. The authoritative set is always derived server-side from the referenced assets, so: - `content_types` is optional. When supplied it is validated against the derivation and must agree with it, in any order. Omit it and the server derives it silently. - `content_type` (singular) is **deprecated** and kept for older clients. When supplied it asserts that every asset shares that one type, and a mixed selection is rejected. Send `null` or omit it and use `content_types` instead. **Communication type.** Must be drawn from the list matching the evaluation's content types — see the `settings` endpoint. A single-type evaluation uses that type's list; an evaluation spanning more than one type must use one of the *mixed* communication types, and the two sets are mutually exclusive. **Model support.** `source` must support every content type in the evaluation. A selection containing a video therefore restricts the choice to the video-capable models. **Limits.** 1–10 assets, of which at most 4 may be videos (videos are the expensive modality and stay capped even when combined with other types). 3–6 attributes. Creating an evaluation consumes workspace quota and its cost in generations is priced per asset by that asset's own type — see the `compute-cost` endpoint to price a selection before committing to it. Required permission: `write:asset_evaluation`. ## OpenAPI ````yaml https://openapi.shareofmodel.ai/swagger.json post /v1/organizations/{organization_id}/workspaces/{workspace_id}/asset_evaluations openapi: 3.0.3 info: title: Share Of Model API version: v1 description: >- ## Model Context Protocol (MCP) In addition to this REST API, Share of Model exposes a **Model Context Protocol** server that lets AI assistants (Claude Desktop, Claude Code, MCP Inspector, custom agents…) call our endpoints directly as tools. Any MCP-compatible client can interact with Share of Model without writing custom integration code — connect once with your usual login and start asking the assistant to query the data for you. ### Connecting from Claude Desktop Open **Settings → Connectors**, scroll to the bottom and click **Add custom connector**, then paste `https://mcp.shareofmodel.ai/mcp/`. A browser window opens for you to log in with your Share of Model account (same login as the web app), and the assistant gains access to the tools. ### Connecting from Claude Code ```bash claude mcp add --transport http share-of-model https://mcp.shareofmodel.ai/mcp/ ``` The first time you call a tool, Claude Code opens your browser to complete the login. ### Connecting from MCP Inspector ```bash npx @modelcontextprotocol/inspector ``` In the Inspector UI, pick **Streamable HTTP** as transport, paste `https://mcp.shareofmodel.ai/mcp/`, and click **Connect**. The first connection prompts you to log in. ### Available tools Only endpoints tagged `mcp` in this OpenAPI spec are exposed as MCP tools, and only read-only (`GET`) routes are exposed. Everything tagged `mcp` below is callable from any compliant MCP client. ### Example prompts Once connected, try asking your assistant things like: - _"List the workspaces I have access to."_ - _"Show me the latest searches in workspace X."_ - _"Compare the share of model between brand A and brand B over the last 30 days."_ For more details on the protocol itself, see the [Model Context Protocol specification](https://modelcontextprotocol.io/). servers: - description: Production API url: https://api.shareofmodel.ai/ - description: Development API url: https://api.dev.shareofmodel.ai/ security: [] tags: - name: Auth description: Endpoints needed for API authentication. - name: Organizations description: Endpoints related to organizations, to list all available organizations. - name: Workspaces description: Endpoints related to workspaces, to list all available workspaces. - name: Analyses description: Endpoints related to analyses and analyses management. - name: Asset Evaluations description: Endpoints related to assets and asset evaluations. - name: Brand Catalog description: Endpoints related to general brand information. - name: Content Briefs description: Endpoints related to content briefs generation and optimisation. - name: Metrics description: >+ Endpoints related to brand metrics. **LEXICON** **Brand Awareness**: What opinion the LLMs have concerning specific brands, related to certain categories. **Brand Perception**: The general sentiment of the LLMs towards a brand, based on the pros and cons they mention. paths: /v1/organizations/{organization_id}/workspaces/{workspace_id}/asset_evaluations: post: tags: - Asset Evaluations summary: Create asset evaluation description: >- Create a new asset evaluation over assets already in the workspace's library. The assets are referenced by `assets_ids` and must all belong to the workspace in the path; referencing an asset from another workspace is rejected with `400`. **Content types.** An evaluation may span more than one content type — for example an image and a video in the same evaluation. The authoritative set is always derived server-side from the referenced assets, so: - `content_types` is optional. When supplied it is validated against the derivation and must agree with it, in any order. Omit it and the server derives it silently. - `content_type` (singular) is **deprecated** and kept for older clients. When supplied it asserts that every asset shares that one type, and a mixed selection is rejected. Send `null` or omit it and use `content_types` instead. **Communication type.** Must be drawn from the list matching the evaluation's content types — see the `settings` endpoint. A single-type evaluation uses that type's list; an evaluation spanning more than one type must use one of the *mixed* communication types, and the two sets are mutually exclusive. **Model support.** `source` must support every content type in the evaluation. A selection containing a video therefore restricts the choice to the video-capable models. **Limits.** 1–10 assets, of which at most 4 may be videos (videos are the expensive modality and stay capped even when combined with other types). 3–6 attributes. Creating an evaluation consumes workspace quota and its cost in generations is priced per asset by that asset's own type — see the `compute-cost` endpoint to price a selection before committing to it. Required permission: `write:asset_evaluation`. operationId: create_asset_evaluation parameters: - in: path name: organization_id schema: type: string format: uuid description: A UUID string identifying the organization. required: true - in: path name: workspace_id schema: type: string format: uuid description: A UUID string identifying the workspace. required: true requestBody: content: application/json: schema: $ref: '#/components/schemas/CreateAssetEvaluationRequest' application/x-www-form-urlencoded: schema: $ref: '#/components/schemas/CreateAssetEvaluationRequest' multipart/form-data: schema: $ref: '#/components/schemas/CreateAssetEvaluationRequest' '*/*': schema: $ref: '#/components/schemas/CreateAssetEvaluationRequest' required: true responses: '201': content: application/json: schema: $ref: '#/components/schemas/AssetEvaluation' description: '' '400': description: >- Invalid input: assets missing from the workspace, `content_types` disagreeing with the referenced assets, a communication type that does not match the content types, a model that does not support every content type, or too many video assets. '403': description: The workspace's asset evaluation quota is fully used. '503': description: The evaluation could not be created downstream. Safe to retry. security: - Bearer: [] components: schemas: CreateAssetEvaluationRequest: type: object description: Base for all the serializers in the assets app properties: name: type: string source: enum: - gpt-4o - gpt-5 - gpt-5.2 - gpt-5.4-mini - gemini-2.0-flash - gemini-2.5-flash - gemini-3-flash-preview - meta-llama-3.3-70B-instruct-turbo - meta-llama-4-maverick - claude-4-6-sonnet type: string x-spec-enum-id: 156d86aecf288cb7 description: >- AI model that should perform the evaluation * `gpt-4o` - gpt 4o * `gpt-5` - gpt 5 * `gpt-5.2` - gpt 5.2 * `gpt-5.4-mini` - gpt 5.4 mini * `gemini-2.0-flash` - gemini 2.0 flash * `gemini-2.5-flash` - gemini 2.5 flash * `gemini-3-flash-preview` - gemini 3 flash preview * `meta-llama-3.3-70B-instruct-turbo` - meta llama 3.3 70B instruct turbo * `meta-llama-4-maverick` - meta llama 4 maverick * `claude-4-6-sonnet` - claude 4 6 sonnet brand: type: string category: type: string description: >- The category that best describes the type of product or service the brand offers attributes: type: array items: type: string maxItems: 6 minItems: 3 assets_ids: type: array items: type: string format: uuid maxItems: 10 minItems: 1 communication_type: enum: - Brand Taglines - Social Media Captions - Campaign Messaging - Email Campaign Copy - Landing Page Headlines - Product Descriptions - Blog Post - Press Release - Other Text Category - Campaign Visuals - Social Media Ads - Website Page - Product Feature Highlights - Display Banners - Not Specified Image - Social Media Shorts - Customer Testimonial - Campaign Ads - YouTube Video - User-Generated Content - Other Video Category - Campaign Materials - Social Media Content - Advertising & Promotions - Product & Feature Highlights - Website & Landing Pages - Customer & User Content - Editorial & Informational Content - Branding & Identity type: string x-spec-enum-id: e5b3ccb79fbae892 description: >- Type of communication the asset is related to. An evaluation whose assets span more than one content type must use one of the mixed communication types. * `Brand Taglines` - Brand Taglines * `Social Media Captions` - Social Media Captions * `Campaign Messaging` - Campaign Messaging * `Email Campaign Copy` - Email Campaign Copy * `Landing Page Headlines` - Landing Page Headlines * `Product Descriptions` - Product Descriptions * `Blog Post` - Blog Post * `Press Release` - Press Release * `Other Text Category` - Other Text Category * `Campaign Visuals` - Campaign Visuals * `Social Media Ads` - Social Media Ads * `Website Page` - Website Page * `Product Feature Highlights` - Product Feature Highlights * `Display Banners` - Display Banners * `Not Specified Image` - Not Specified Image * `Social Media Shorts` - Social Media Shorts * `Customer Testimonial` - Customer Testimonial * `Campaign Ads` - Campaign Ads * `YouTube Video` - YouTube Video * `User-Generated Content` - User Generated Content * `Other Video Category` - Other Video Category * `Campaign Materials` - Campaign Materials * `Social Media Content` - Social Media Content * `Advertising & Promotions` - Advertising & Promotions * `Product & Feature Highlights` - Product & Feature Highlights * `Website & Landing Pages` - Website & Landing Pages * `Customer & User Content` - Customer & User Content * `Editorial & Informational Content` - Editorial & Informational Content * `Branding & Identity` - Branding & Identity content_type: enum: - image - text - video - null type: string x-spec-enum-id: 0343cca47d7cc465 nullable: true description: >- DEPRECATED — superseded by `content_types`. When provided, every asset must match it. * `image` - image * `text` - text * `video` - video content_types: type: array items: enum: - image - text - video type: string description: |- * `image` - image * `text` - text * `video` - video x-spec-enum-id: 0343cca47d7cc465 description: >- The distinct content types of the referenced assets, e.g. ['image', 'video']. Derived server-side from `assets_ids`; when supplied it is validated against that derivation. persona: type: string description: >- Persona considered for the evaluation, example: '{age} {gender} {interest or habits}' description: type: string callback_url: type: string description: URL to send the evaluation result to tags: type: array items: type: string maxLength: 255 description: Optional list of tags for the asset evaluation required: - assets_ids - attributes - brand - category - communication_type - name - source AssetEvaluation: type: object properties: id: type: string format: uuid readOnly: true content_types: type: array items: enum: - image - text - video type: string description: |- * `image` - image * `text` - text * `video` - video x-spec-enum-id: 0343cca47d7cc465 readOnly: true created_at: type: string format: date-time readOnly: true title: Created created_by: type: string format: uuid workspace_id: type: string format: uuid status: enum: - pending - in_progress - completed - partial - failed type: string description: |- * `pending` - pending * `in_progress` - in progress * `completed` - completed * `partial` - partial * `failed` - failed x-spec-enum-id: 07c86065df656b0c name: type: string maxLength: 255 source: enum: - gpt-3.5-turbo - gpt-4-turbo - gpt-4o - gpt-4o-mini-search - gpt-5 - gpt-5.2 - gpt-5.4-mini - gemini-1.5-pro - gemini-2.0-flash - gemini-2.5-flash-grounded - gemini-2.5-flash - gemini-2.5-flash-lite - gemini-3-flash-preview - meta-llama-3.1-70B-instruct-turbo - meta-llama-3.2-70B-instruct-turbo - meta-llama-3.3-70B-instruct-turbo - meta-llama-4-maverick - mistral-7B-instruct - claude-3-5-sonnet - claude-4-sonnet - claude-4-5-sonnet - claude-4-6-sonnet - deepseek-chat - deepseek-reasoner - deepseek-v3.2-maas - perplexity-sonar - rufus - google-ai-mode type: string description: >- * `gpt-3.5-turbo` - gpt 3.5 turbo * `gpt-4-turbo` - gpt 4 turbo * `gpt-4o` - gpt 4o * `gpt-4o-mini-search` - gpt 4o mini search * `gpt-5` - gpt 5 * `gpt-5.2` - gpt 5.2 * `gpt-5.4-mini` - gpt 5.4 mini * `gemini-1.5-pro` - gemini 1.5 pro * `gemini-2.0-flash` - gemini 2.0 flash * `gemini-2.5-flash-grounded` - gemini 2.5 flash grounded * `gemini-2.5-flash` - gemini 2.5 flash * `gemini-2.5-flash-lite` - gemini 2.5 flash lite * `gemini-3-flash-preview` - gemini 3 flash preview * `meta-llama-3.1-70B-instruct-turbo` - meta llama 3.1 70B instruct turbo * `meta-llama-3.2-70B-instruct-turbo` - meta llama 3.2 70B instruct turbo * `meta-llama-3.3-70B-instruct-turbo` - meta llama 3.3 70B instruct turbo * `meta-llama-4-maverick` - meta llama 4 maverick * `mistral-7B-instruct` - mistral 7B instruct * `claude-3-5-sonnet` - claude 3 5 sonnet * `claude-4-sonnet` - claude 4 sonnet * `claude-4-5-sonnet` - claude 4 5 sonnet * `claude-4-6-sonnet` - claude 4 6 sonnet * `deepseek-chat` - deepseek chat * `deepseek-reasoner` - deepseek reasoner * `deepseek-v3.2-maas` - deepseek v3.2 maas * `perplexity-sonar` - perplexity sonar * `rufus` - rufus * `google-ai-mode` - google ai mode x-spec-enum-id: 31229724ccbe9afa brand: type: string maxLength: 255 category: type: string maxLength: 255 persona: type: string maxLength: 255 attributes: type: array items: type: string maxLength: 255 description: type: string is_archived: type: boolean communication_type: type: string maxLength: 255 tags: type: array items: type: string maxLength: 255 generation_cost: type: integer maximum: 2147483647 minimum: -2147483648 required: - brand - category - content_types - created_at - created_by - id - name - persona - source - workspace_id securitySchemes: Bearer: type: apiKey in: header name: Authorization ````