Picsha AI MCP Server
OfficialServer Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PICSHA_API_TOKEN | Yes | Your Picsha AI API token, generated from the Picsha Admin Dashboard. Grants the AI agent access to your organization's library. | |
| PICSHA_EXTERNAL_USER_ID | No | Optional. Dynamically restricts the agent's context to a specific user by injecting their User ID. Used for sandbox mode or user isolation. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
| prompts | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| search_assetsA | Search for assets in the Picsha AI platform using vector or standard keyword search. Note: If your agent instance is sandboxed to a specific user via environment variables, this search will ONLY return assets owned by that specific user. You are securely retrieving their contextual assets. |
| get_assetB | Retrieve detailed metadata and AI analysis results for a specific asset |
| reanalyze_assetB | Manually trigger AI re-analysis on an existing asset to detect faces, objects, etc. |
| summarize_assetB | Utilize Claude Sonnet via Amazon Bedrock to summarize documents on demand |
| get_rendered_asset_urlB | Generate a dynamic delivery URL for an asset with transformation parameters (e.g. width, height, format, smart crop) |
| upload_assetA | Upload a local file directly to the Picsha AI platform. This acts as a proxy, fetching a pre-signed S3 URL and executing the PUT request automatically. If your agent is running with user sandboxing, this file will automatically be securely bound to that user's identity. Note: Uploading immediately triggers the asynchronous 'picsha-ai-ingest' pipeline which will extract metadata, generate thumbnails, and run AI analysis (faces, tags, bedrcock summaries). Therefore, the returned asset will initially be in a 'pending' state. You should use the 'get_asset' tool a few seconds after uploading to retrieve the final AI-processed results. |
| list_recent_assetsB | List recently added assets in the Picsha platform. IMPORTANT: When replying to the user, ALWAYS format this as a clean Markdown table with columns for ID, Original Name, Status, and Date. |
| update_assetA | Update an asset's tags or custom metadata. Prefix a tag with a hyphen (e.g. '-discard') to remove it, or specify normally to add it. |
| delete_assetA | Permanently delete an asset from the database, search indexes, and physical storage. |
| moderate_assetB | Approve or reject a moderated asset pending manual review. |
| create_dam_groupA | Create a new Digital Asset Management (DAM) group/collection (folder) to organize assets. |
| link_assetsB | Link a source/parent asset to a target/child asset (e.g. variations, derived formats, social crops) with a custom relationship description. |
| trigger_url_ingestB | Ingest a public web asset directly into the Picsha AI platform by downloading and putting it through the ingestion pipeline. |
| escalate_to_supportA | Use this tool ONLY when you need to log a feature request, report a documentation gap, or escalate an issue to the engineering team. This will actually send an email to support@picsha.ai. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| analyze_asset_profile | Provides a structured template to do a deep analysis of a media asset's metadata, EXIF details, and AI tags. |
| generate_social_campaign | Helps generate platform-specific social media copy and smart crop parameters based on an asset's content. |
| image_magic_transform | Walks the LLM and user through Picsha's generative AI fill and background removal parameters. |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 14 tools
Each tool has a clearly distinct purpose: CRUD for assets, group creation, search, ingestion, AI analysis, moderation, and support escalation. Overlap between upload_asset and trigger_url_ingest is clarified by their descriptions (local file vs URL).
Tool names uniformly follow snake_case with a verb_noun pattern (e.g., create_dam_group, get_asset, update_asset). This provides predictable naming for an agent to infer functionality.
14 tools is well-scoped for a digital asset management platform with AI features. Each tool represents a distinct action without unnecessary redundancy, covering ingestion, search, retrieval, analysis, and administration.
Core asset lifecycle (create/read/update/delete), search, AI analysis, and moderation are covered. Missing update/deletion for groups and lack of bulk operations are minor gaps, but the surface is solid for typical workflows.