GuidesUnderstand DAM
AI, MCP and Content Credentials in DAM: which products ship what
Which DAM products document AI search, an MCP server for AI assistants or Content Credentials (C2PA), plus an EU AI Act Article 50 checklist.
- 3
- capabilities compared
- <1%
- of images online carry Content Credentials
On this page(6 sections)
AI search, access for AI assistants and proof of where an image came from are the newest capabilities in digital asset management, and the ones that change fastest. Vendors announce them well before they document them, so a buyer comparing products has to separate what ships today from what is promised.
This article shows which DAM products document each capability, explains MCP and Content Credentials in plain terms, and gives a checklist for the EU AI Act transparency rules that apply to marketing teams. Use it to check a shortlist against what you need, and to know what to ask a vendor.
As of 2026-10-19, few DAM vendors visibly ship all three. Acquia DAM documents an MCP server (read-only first release, August 2026), Air lists one on its Business plan, Fotoware announced end-to-end Content Credentials support in September 2026, and Freedam documents all three: AI on every plan, an MCP server and Content Credentials reading and filtering. Fewer than 1% of images online carry Content Credentials data.
Which DAM products ship AI, MCP or Content Credentials?#
| Product | What is documented |
|---|---|
| Freedam (publisher) | AI included on every plan: tagging, captions, OCR (text recognition), transcription, semantic search, and face recognition with human-confirmed names (AI guide). Content Credentials reading and filtering (Content Credentials). An MCP server (MCP for developers) |
| Acquia DAM | MCP server: assistants such as Claude, Microsoft Copilot Studio and ChatGPT can search assets and read asset details and metadata, with OAuth (a sign-in standard) and per-user permissions. Read-only first release; enabled through Acquia support for customers with the REST API (Acquia documentation) |
| Fotoware | End-to-end C2PA support; search filters exclude files with invalid credentials, AI-processed files and files not captured by a real camera (Fotoware) |
| TESSA DAM | Automatic AI Generated and AI Modified labels stored in the asset metadata, with an AI-Maker filter to search them (TESSA) |
| Air | MCP server listed on the Business plan (Air) |
| Aprimo | Published explainer on MCP for DAM agents (Aprimo blog); product status not stated there |
| Adobe (AEM Cloud Service) | Search by intent and meaning across about 100 languages, using embeddings; a Brand Governance Agent that scans for expired, duplicate or non-compliant assets |
What is MCP, and why does a DAM need it?#
The Model Context Protocol (MCP) is a standard that lets an AI assistant use another system as a tool. For a DAM, an MCP server lets an assistant search the library, read an asset's details and metadata and, in later releases, act on them. Acquia's first release is read-only and uses OAuth with per-user permissions, so the assistant sees only what that user may see.
What to test: whether the assistant respects each user's permissions, whether it can write as well as read, and whether its actions appear in the audit log.
What do buyers want AI to do in a DAM?#
In MediaValet's DAM Trends Report 2026:
| Use of AI | |
|---|---|
| Asset discovery | 82% |
| Curation | 73% |
| Keeping content on-brand | 63% |
| Generation or editing | 30% |
G2 reports vendors estimating AI feature usage from 11 to 25% up to above 75% of their customers, so adoption is uneven.
What are Content Credentials, and which DAMs support them?#
Content Credentials are provenance data from the C2PA (Coalition for Content Provenance and Authenticity) standard, attached to a file: who made it, with what tool and what changed. C2PA version 2.3 shipped in February 2026 and is being fast-tracked as an ISO (International Organization for Standardization) standard. Fewer than 1% of images online carry C2PA data. Competitor coverage is thin: only a few vendors visibly ship it, and Fotoware's announcement is the most specific we found. See the Content Credentials guide for what a DAM should do with them.
What a DAM should do with Content Credentials: read them on upload, keep them when files are exported, and let you filter the library by them.
What does EU AI Act Article 50 require of marketing teams?#
Transparency obligations apply from 2 August 2026. Realistic AI-generated or manipulated images, audio and video, and certain AI-generated text on matters of public interest, need disclosure. Advertisers, agencies and organizations outside the EU whose content reaches EU audiences can be in scope. The European Commission published guidelines on 20 July 2026 (a summary of the guidance, not legal advice).
A checklist for the asset library:
- Can you tell which assets were AI-generated or AI-edited? Store it as a metadata field.
- Is the origin recorded at upload (file credentials, creator field or workflow)?
- Can you filter the library for AI-generated assets before publishing?
- Is the label applied where the asset is published, not only in the DAM?
- Does the DAM keep credentials on export, or does it strip them?
- Who signs off on the disclosure, and is that recorded?
Ask counsel before relying on this list.
When a larger vendor is the better fit: if you need AI inside a full enterprise suite, with workflow, rights and governance modules from one supplier, Adobe, Aprimo or Acquia may serve you better than a smaller product with fewer modules.
Last verified: 2026-10-19.
Frequently asked questions#
Acquia DAM documents an MCP server, read-only in its first release (August 2026). Freedam offers an MCP server. Air lists one on its Business plan, and Aprimo has published an explainer.
Fotoware announced end-to-end C2PA support in September 2026, with filters for files with invalid credentials. Freedam reads and filters on Content Credentials. We found few others in public documentation.
From 2 August 2026, realistic AI-generated or manipulated images, audio and video need disclosure. Your library should record which assets are AI-generated and keep that information with the asset. Check with counsel.
Auto-tagging attaches labels to assets when they are uploaded. Semantic search matches the meaning of a query to the content of assets, so it finds results even when no tag matches the words typed.
Not to be a DAM. AI changes who writes the first draft of the metadata. A small, well-tagged library works without it; a library of tens of thousands of untagged files is where it saves the most time.
Found a wrong figure or a missing product? Report an error with a source. How the data was collected is in the method section.