The AI Media Index

Supported actions

Read

Yes

Create

No

Modify

No

Delete

No

Account discovery, metadata and GAQL queries. Campaign building and optimisation still happen elsewhere.

Why this grade

• Official, maintained MCP with a clear read boundary.
• Google Ads Query Language supports detailed account and performance diagnosis.
• Useful as a reporting and second-pair-of-eyes surface once configured.

Limits

• The MCP cannot create, modify or delete any campaign object.
• Setup requires Python, a developer token, Google Cloud, enabled APIs, OAuth and client configuration.
• The workflow stops at diagnosis and cannot perform the optimisation it recommends.

Check the sources

Supported actions

Read

Yes

Create

Partial

Modify

No

Delete

No

Creation and analysis are present, but the current beta does not provide dependable general modification of existing campaign objects.

Why this grade

• Hosted OAuth makes the connector much easier to add than a custom Marketing API integration.
• It can read performance and campaign structure in the assistant.
• It can create supported campaign, ad-set and ad objects.

Limits

• The current beta does not provide dependable general modification of existing campaign objects.
• Availability still varies by ad account and by individual tool.
• Creation without ordinary follow-up editing sends the operator back to Ads Manager.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Around 400 tools cover creation, targeting, bidding, reporting and creative operations. Some manual and account workflows remain outside the route.

Why this grade

• Official remote MCP covers campaign, reporting, audience and creative work.
• Read/write capability makes it useful beyond analysis.
• The interface is designed for advertiser workflows rather than a generic API bridge.

Limits

• Broad write permissions increase the cost of a bad instruction.
• Business account eligibility and permissions still govern availability.
• No dedicated official CLI was verified.

Check the sources

Supported actions

Read

Yes

Create

No

Modify

No

Delete

No

Current launch documentation covers exploration, performance queries and audits, not mutation.

Why this grade

• Official managed connection is easier than building against the API.
• Provides entity-level performance data for campaigns, ad groups, ads and keywords.
• Useful for conversational audits and cross-platform reporting.

Limits

• The documented launch surface is read-only.
• It cannot create an ad or apply the campaign changes suggested by an audit.
• Broad manage-campaign language on the product page overstates the action boundary described in its FAQ.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Yes

Includes individual API capabilities plus tools for account creation, reports, campaign launch and locale expansion.

Why this grade

• Official MCP can create, update and delete campaign objects.
• Retail-media context makes it strategically important.
• Uses the established Amazon Ads API permission model.

Limits

• The MCP is still an open beta.
• Access is limited to partners with active API credentials.
• Destructive campaign operations require strong human approval gates.

Check the sources

Supported actions

Read

Yes

Create

No

Modify

No

Delete

No

Retrieval and analysis only. The broader Marketing API is a separate engineering route.

Why this grade

• Official hosted MCP provides campaign retrieval and analysis.
• Read-only control is suitable for a cautious reporting pilot.
• The underlying Marketing API offers a route to deeper integration.

Limits

• MCP access is still allowlisted.
• The MCP cannot execute campaign changes.
• No official plugin or dedicated CLI was verified.

Check the sources

Supported actions

Read

Yes

Create

Partial

Modify

Partial

Delete

Partial

The API covers important campaign-management jobs, but access is approved by product and use case and there is no official agent interface.

Why this grade

• Official Marketing APIs cover campaign management and reporting.
• Strong fit for B2B audience and lead-generation workflows.
• Production access tiers and permissions are documented.

Limits

• No official MCP was verified.
• API product approval and access tiers add friction.
• Daily AI use needs a custom or third-party integration layer.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Useful campaign and reporting coverage through the API, without a vendor-backed MCP, plugin or CLI.

Why this grade

• Official Ads API supports campaign and reporting workflows.
• Useful specialist channel for visual discovery and commerce.
• An authorised integration can run recurring production tasks.

Limits

• No official MCP or agent plugin was verified.
• No dedicated official CLI was verified.
• AI workflows require custom engineering or a maintained intermediary.

Check the sources

Supported actions

Read

Yes

Create

Partial

Modify

Partial

Delete

Partial

Core campaign operations exist in the API, but no official agent interface was verified.

Why this grade

• Official Ads API v3 exposes campaign and account operations.
• Community context can be valuable for specific audience strategies.
• OAuth provides a conventional route for controlled integrations.

Limits

• No official MCP, plugin or CLI was verified.
• Advertising API access requires approval.
• A useful daily agent workflow must be built and maintained separately.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Read accounts and results, create and update campaigns and line items, manage targeting and supported creative and audience objects.

Why this grade

• Official Ads-specific MCP documents campaign and line-item writes.
• New campaigns and line items start paused, with separate activation tools.
• Includes targeting, creative, audience and conversion-tracking tools.

Limits

• Developer-app and Ads access remain prerequisites.
• Campaign deletion is not in the documented MCP tool list.
• Activation and budget changes require explicit operating controls.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Broad trafficking coverage, but performance reporting is separate and some YouTube and Partners writes are unsupported.

Why this grade

• Mature official API covers core programmatic campaign resources.
• Supports insertion orders, line items, targeting, creatives and users.
• Role-based access can separate read, standard and admin operations.

Limits

• No official MCP or agent plugin was verified.
• Performance reporting requires the separate Bid Manager API.
• Some YouTube and Partners write workflows remain outside the API.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Strong trafficking and measurement coverage, but no official agent interface packages the resource model for daily use.

Why this grade

• Mature API supports trafficking, creative, conversion and reporting workflows.
• Broad resource coverage suits complex enterprise operations.
• User-profile permissions remain the control plane.

Limits

• No official MCP, plugin or dedicated CLI was verified.
• The resource model and permissions create implementation overhead.
• Write access can affect live trafficking and needs a second-person check.

Check the sources

Supported actions

Read

Yes

Create

No

Modify

No

Delete

No

High-volume reporting only. Offline conversion upload is a separate surface and ordinary campaign writes are absent.

Why this grade

• High-volume official reporting API with a purpose-built query language.
• Immediate access to granular campaign, ad and keyword data.
• REST, gRPC and official client libraries support production pipelines.

Limits

• The reporting API is read-only.
• Offline conversions use a separate API surface.
• No official MCP or dedicated agent integration was verified.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Core campaign operations are covered, but no official MCP, agent plugin or daily CLI was verified.

Why this grade

• Official API supports campaign, ad-group, keyword and reporting work.
• API roles distinguish read-only from read/write access.
• Covers ads on the App Store and Apple Maps.

Limits

• No official MCP, plugin or dedicated CLI was verified.
• Third-party providers require separate registration.
• Daily AI use needs a custom integration and careful OAuth setup.

Check the sources

Supported actions

Read

Yes

Create

Yes

Modify

Yes

Delete

Partial

Published scope spans data and actions, but exact tool coverage is not public and access is private beta.

Why this grade

• Official OAK MCP connects AI assistants to Ask Koa.
• Published scope includes querying data, creating campaigns and monitoring live activity.
• The underlying REST and GraphQL APIs cover substantial platform operations.

Limits

• OAK is available only to selected partners in private beta.
• Broad action scope makes approval and audit controls essential.
• Practical onboarding still depends on an account representative.

Check the sources

Supported actions

Context

Yes

Create

Yes

Edit

Partial

Adapt

Partial

Export

Yes

Useful design creation and editing, without full parity with Canva's manual editor and app ecosystem.

Why this grade

• Official MCP supports design creation, editing, assets and export.
• Remote OAuth makes it accessible to common agent clients.
• Apps CLI and Connect APIs support deeper product integration.

Limits

• The Apps CLI is not a daily creative-production CLI.
• Brand and format governance still need human review.
• It does not validate media-platform setup or campaign delivery.

Check the sources

Supported actions

Context

Yes

Create

Partial

Edit

Partial

Adapt

Partial

Export

Partial

Excellent design context and code handoff. Write-to-canvas is useful but does not reproduce the full editor or plugin ecosystem.

Why this grade

• Official remote and desktop MCP expose real design context.
• Supported clients can create or modify native canvas content.
• Plugin API and agent skills support structured creative workflows.

Limits

• Write-to-canvas is beta and depends on current seat eligibility and file permissions.
• It does not connect creative output to media-platform delivery by itself.
• No dedicated official Figma CLI was verified.

Check the sources

Supported actions

Context

Yes

Create

Yes

Edit

Yes

Adapt

Yes

Export

Yes

Can execute and monitor multi-step production workflows, with human brand, legal and final-quality approval.

Why this grade

• Hosted MCP runs real reusable creative workflows.
• Supports generation, editing, composition and format variants.
• Asynchronous monitoring fits large production jobs.

Limits

• Enterprise-only access limits practical adoption.
• Workflow design must exist before reliable reuse.
• Human brand and output review remains essential.

Check the sources

Supported actions

Context

Partial

Create

Yes

Edit

Yes

Adapt

Yes

Export

Yes

Creates and transforms production media across modalities; the brief, brand system, rights review and final QC remain human responsibilities.

Why this grade

• Official MCP, API, SDKs and open GitHub server.
• Creates and edits image, video and audio assets.
• Task APIs and reusable workflows support automation.

Limits

• Outputs still need rights and brand review.
• Temporary result URLs require disciplined download.
• Generation cost and model choice need controls.

Check the sources

Supported actions

Analyse

Yes

Build Reports

Partial

Configure

No

Validate

Partial

Share

Partial

Strong querying and reporting, but events, audiences and property settings cannot be changed through the MCP.

Why this grade

• Official open-source MCP is suitable for recurring reporting and analysis.
• Read-only design limits accidental property changes.
• Google Analytics data and administration APIs provide a mature foundation.

Limits

• The MCP cannot edit events, audiences or property settings.
• Cloud credentials and property permissions still require setup.
• No separate official plugin or marketer-facing CLI was verified.

Check the sources

Supported actions

Analyse

Yes

Build Reports

Yes

Configure

Yes

Validate

Partial

Share

Yes

Strong analytics-asset workflow. Raw event instrumentation and ingestion remain outside the MCP.

Why this grade

• Official MCP supports both analysis and write operations.
• Official agent plugins, skills and a CLI create several usable entry points.
• Strong fit for behavioural analysis and experimentation workflows.

Limits

• Write capability requires careful project and taxonomy governance.
• It is an analytics layer, not a media-buying control plane.
• Teams still need consistent event instrumentation for useful answers.

Check the sources

Supported actions

Analyse

Yes

Build Reports

Yes

Configure

Partial

Validate

Partial

Share

Unverified

Query analytics, create and update dashboards, edit Lexicon metadata and manage supported cohorts and other assets.

Why this grade

• The official tool reference now documents dashboard creation, updates and deletion.
• Lexicon editing and data-quality tools extend the route beyond reporting.
• Regional endpoints and OAuth or service-account authentication are documented.

Limits

• Experiments, feature flags and audit-log tools have beta or access limits.
• Free-plan query limits and organisation enablement can affect availability.
• Documented tools still need a controlled account-level pilot.

Check the sources

Supported actions

Analyse

Yes

Build Reports

Yes

Configure

Yes

Validate

Partial

Share

Partial

It completes ad-hoc analysis and creates reusable components or projects; governed interpretation and stakeholder delivery remain human work.

Why this grade

• Hosted official MCP with normal Adobe permissions.
• Creates and runs ranked or trended reports.
• Can manage segments, date ranges and workspace projects.

Limits

• Enterprise entitlement and administrator permissions are required.
• It does not operate ad platforms.
• Report correctness still depends on governed suites and components.

Check the sources

Supported actions

Analyse

Yes

Build Reports

Yes

Configure

Yes

Validate

Partial

Share

Partial

Strong analysis and component coverage once the cross-channel data model exists; upstream data design stays outside the route.

Why this grade

• Official hosted MCP for cross-channel analysis.
• Uses the organisation existing CJA permissions.
• Can work with reusable components and projects.

Limits

• Useful only after substantial data-view and identity work.
• Enterprise access raises the setup bar.
• It cannot repair upstream instrumentation.

Check the sources

Supported actions

Analyse

Yes

Build Reports

Yes

Configure

Yes

Validate

Yes

Share

Partial

The route can investigate behaviour and maintain reusable insights and dashboards; distribution and paid-media decisions remain outside it.

Why this grade

• Queries the major product-analytics methods and SQL.
• Creates, updates and deletes insights and dashboards.
• One-command wizard installation is unusually direct.

Limits

• It is product-led rather than paid-media-led.
• Quality depends on the event implementation.
• It does not control media spend or delivery.

Check the sources

Supported actions

Inspect

Yes

Configure

Yes

Validate

Partial

Deploy

Yes

Monitor

Partial

Most configuration objects are available through the API, including version and publish workflows.

Why this grade

• Official API supports container, workspace, tag, trigger and variable workflows.
• Useful for repeatable measurement deployment and governance.
• Workspace and version concepts support controlled changes.

Limits

• No official MCP, agent plugin or dedicated CLI was verified.
• Incorrect writes can damage live measurement.
• Safe automation needs preview, review and publication gates outside the API.

Check the sources

Supported actions

Inspect

Yes

Configure

No

Validate

Partial

Deploy

No

Monitor

Yes

Useful performance monitoring and implementation diagnosis, without a dependable write route for correcting configuration.

Why this grade

• Hosted official MCP with OAuth and token routes.
• Useful campaign and attribution diagnosis in chat.
• Includes implementation and OneLink troubleshooting.

Limits

• The service is explicitly beta.
• Sessions can require frequent reconnection.
• The reviewed route diagnoses more than it configures.

Check the sources

Supported actions

Inspect

Yes

Configure

Partial

Validate

Partial

Deploy

Partial

Monitor

Yes

Broad developer capability exists, but it is fragmented and not presented as a ready AI route.

Why this grade

• Mature APIs and SDKs cover measurement and reporting.
• App Automation can create or update configured apps.
• Device APIs support testing workflows.

Limits

• No official MCP or agent plugin was verified.
• Capabilities are split across several APIs.
• Some automation requires vendor enablement.

Check the sources

Supported actions

Inspect

Yes

Configure

Yes

Validate

Partial

Deploy

Yes

Monitor

Partial

The Public API can manage core workspace resources, but event design and quality validation still require specialist work.

Why this grade

• Mature collection and routing infrastructure.
• Public API provides CRUD for core workspace resources.
• Strong governance surface for tracking plans and destinations.

Limits

• No official MCP was verified.
• A custom agent layer is required.
• Identity and instrumentation design remain specialist work.

Check the sources

Supported actions

Discover

Yes

Retrieve

Yes

Compare

Yes

Analyse

Partial

Export

Yes

Wide research-data coverage. It does not read or control the advertiser's own paid-media accounts.

Why this grade

• Official MCP and CLI make search and competitive data directly usable by agents.
• Read-only control is appropriate for external research data.
• Broad API coverage supports repeatable market and SERP analysis.

Limits

• It does not operate advertising accounts.
• Usage-based API costs need monitoring.
• Raw results still require paid-media judgement before action.

Check the sources

Supported actions

Discover

Yes

Retrieve

Yes

Compare

Yes

Analyse

Yes

Export

Partial

Strong search and competitor research without custom integration; decision framing and activation remain outside the product.

Why this grade

• Hosted official MCP with straightforward OAuth.
• Strong keyword, backlink and competitor evidence.
• Useful without building a custom API wrapper.

Limits

• Read-only research rather than activation.
• Requires an eligible paid plan.
• It does not include first-party campaign truth.

Check the sources

Supported actions

Discover

Yes

Retrieve

Yes

Compare

Yes

Analyse

Yes

Export

Yes

Broad report discovery and execution across Semrush datasets; changes to projects or campaigns stay elsewhere.

Why this grade

• Hosted OAuth endpoint works across major clients.
• Broad SEO, paid-search, shopping and market reports.
• Discovery tools make a large API easier for agents.

Limits

• Every call consumes API units.
• Project methods exposed by MCP are read-only.
• Subscription entitlements make coverage variable.

Check the sources

Supported actions

Discover

Yes

Retrieve

Yes

Compare

Yes

Analyse

Yes

Export

Yes

Wide external-market coverage and direct comparisons, bounded by licensed datasets and data-credit economics.

Why this grade

• Hosted MCP covers more than 75 intelligence endpoints.
• Strong web, app, search, audience and retail breadth.
• Supports direct competitor comparison.

Limits

• Requires API-inclusive paid access.
• Queries consume data credits.
• Estimated and delayed datasets are not first-party live truth.

Check the sources

Supported actions

Connect

Yes

Build

Partial

Run

Yes

Debug

Partial

Monitor

Partial

Large app catalogue, but each action must exist in the connector and be explicitly exposed.

Why this grade

• Official MCP reaches a broad catalogue of business applications.
• Read/write actions make it useful for practical cross-tool workflows.
• Platform CLI supports maintained custom integrations.

Limits

• Connector depth and field coverage vary by application.
• Broad action access can create unintended external changes.
• Task limits and commercial plans can constrain high-volume operations.

Check the sources

Supported actions

Connect

Yes

Build

Yes

Run

Yes

Debug

Yes

Monitor

Yes

Broad API and action catalogue, with exact parity determined by the connected component.

Why this grade

• Official MCP and CLI support flexible API-based automation.
• Code steps handle gaps that fixed connectors cannot cover.
• Strong fit for event-driven media and measurement workflows.

Limits

• More engineering-oriented than a typical marketer tool.
• Credential and secret handling needs careful ownership.
• Failures can cross several connected systems and need monitoring.

Check the sources

Supported actions

Connect

Yes

Build

Yes

Run

Yes

Debug

Partial

Monitor

Partial

The documented instance server supports workflow building and editing from n8n 2.13.0, alongside search, testing and execution.

Why this grade

• The instance MCP documents workflow creation and editing as well as execution.
• Workflow enablement and user permissions bound access.
• Self-hosting supports teams that can own the operational setup.

Limits

• Building and editing require n8n 2.13.0 or later.
• Enabled workflow access is shared across connected clients rather than separately selected per client.
• Workflow execution can trigger external writes using connected credentials.

Check the sources

Supported actions

Connect

Yes

Build

Partial

Run

Yes

Debug

Partial

Monitor

Partial

Runs existing scenarios directly and can manage workflow entities on paid plans; complex design still benefits from the visual builder.

Why this grade

• Universal hosted OAuth endpoint.
• Can run scenarios on all plans.
• Paid scopes can manage scenarios and related entities.

Limits

• Scenario quality depends on clear inputs and outputs.
• Management scopes can affect live automation.
• Connector depth still varies by app.

Check the sources

Supported actions

Connect

Partial

Build

Partial

Run

Yes

Debug

Partial

Monitor

Yes

Strong governed execution once recipes and functions are exposed, with coverage determined by enterprise configuration.

Why this grade

• Enterprise-hosted runtime with identity and audit controls.
• Exposes recipes, API collections, functions and skills.
• Strong governance for consequential workflows.

Limits

• Enterprise-only access.
• An administrator must create the endpoint.
• Setup is heavier than self-service competitors.

Check the sources

Supported actions

Connect

Yes

Build

Yes

Run

Yes

Debug

Partial

Monitor

Partial

Strong AI-native workflow construction and execution, with exact depth inherited from each integration and model.

Why this grade

• AI-native agent and workflow construction.
• Supports 50-plus built-in and arbitrary remote MCP servers.
• CLI has JSON output for automation.

Limits

• No universal external Gumloop MCP endpoint was verified.
• Production reliability inherits every connected dependency.
• External MCP servers must be publicly reachable.

Check the sources

Supported actions

Connect

Partial

Inspect

Yes

Query

Yes

Transform

Partial

Operate

Partial

Broad source coverage for retrieval and reporting, but it does not change media-platform objects.

Why this grade

• Official read-only MCP fits recurring cross-channel retrieval.
• Established connector coverage reduces bespoke API maintenance.
• Official AI integrations and CLI provide several workflow options.

Limits

• The CLI remains alpha.
• The MCP cannot change campaigns.
• Coverage and economics depend on the commercial plan and connector.

Check the sources

Supported actions

Connect

Partial

Inspect

Yes

Query

Partial

Transform

Partial

Operate

Partial

Coverage depends on the connected sources and destinations. Documentation search and action-capable MCP surfaces are distinct.

Why this grade

• Official agent MCP can expose useful connector and data-movement context.
• Large connector ecosystem supports cross-channel data collection.
• Open-source and managed routes give teams deployment choice.

Limits

• MCP control is mixed and depends on the configured interface.
• Self-hosting adds operational overhead.
• It is infrastructure rather than a marketer-facing daily interface.

Check the sources

Supported actions

Connect

No

Inspect

Yes

Query

No

Transform

No

Operate

No

The MCP supplies context and retrieval. Connector operations remain in the REST API, SDK or product.

Why this grade

• Mature managed data movement supports reliable recurring pipelines.
• Official MCP, agent plugins and Connector SDK CLI cover several technical workflows.
• Read-only MCP control is a sensible context boundary.

Limits

• Agent Context MCP is still a public preview.
• The MCP is not a substitute for connector operations or warehouse permissions.
• Enterprise cost and administration may be disproportionate for a small stack.

Check the sources

Supported actions

Connect

Depends

Inspect

Depends

Query

Depends

Transform

Depends

Operate

Depends

Prebuilt database tools and custom YAML definitions can expose broad operations, bounded by the configured role and queries.

Why this grade

• Official open-source toolbox exposes governed database tools through MCP.
• Supports BigQuery and other common production data stores.
• CLI and Gemini extensions help teams package repeatable access.

Limits

• It is not a paid-media connector by itself.
• Read/write risk depends entirely on tool definitions and database roles.
• Secure configuration requires data-engineering discipline.

Check the sources

Supported actions

Connect

Depends

Inspect

Yes

Query

Yes

Transform

Depends

Operate

Depends

Coverage can be broad, but it depends entirely on schemas, tools, semantic models and role permissions configured by the team.

Why this grade

• Official managed MCP works inside Snowflake's role and governance model.
• Strong foundation for cross-channel measurement and governed client data.
• Official CLI, APIs and drivers support serious production engineering.

Limits

• It needs warehouse, role and cost governance expertise.
• Mixed control can expose consequential writes when roles are too broad.
• It is infrastructure, not a ready-made paid-media workflow.

Check the sources

Supported actions

Connect

Partial

Inspect

Yes

Query

Partial

Transform

Yes

Operate

Partial

Useful management coverage for customer-data pipelines, with gaps and stability caveats while the server evolves.

Why this grade

• Hosted OAuth MCP is easy to connect.
• Can access and manage sources, destinations and transformations.
• Covers event streams and documentation.

Limits

• Official docs warn that the MCP is still in active development.
• Some product features are missing.
• Connection stability may vary during updates.

Check the sources

Supported actions

Connect

Depends

Inspect

Yes

Query

Yes

Transform

Depends

Operate

Depends

Potentially broad, but every useful marketing workflow depends on the data products and tools an enterprise exposes.

Why this grade

• Managed MCP uses workspace governance and permissions.
• Can expose SQL, Genie and Vector Search.
• Strong fit for governed enterprise data estates.

Limits

• Managed MCP remains public preview.
• Requires substantial platform configuration.
• Not a turn-key paid-media connection.

Check the sources

Supported actions

Connect

Partial

Inspect

Yes

Query

Yes

Transform

Yes

Operate

Yes

Strong analytics-engineering coverage, including build and test, but not a marketer-facing turn-key workflow.

Why this grade

• Official open-source MCP with broad toolsets.
• Strong lineage, health, semantic and SQL context.
• Can run build and test when explicitly enabled.

Limits

• Setup changes materially by toolset.
• Useful operation requires analytics-engineering expertise.
• CLI writes can change warehouse objects.

Check the sources

Supported actions

Prepare

Partial

Configure

Yes

Run

Yes

Validate

Yes

Scenarios

Yes

Covers the MMM lifecycle and scenario planning, not live campaign control.

Why this grade

• Google-backed open-source marketing-mix modelling package.
• Useful for incrementality and budget-allocation questions beyond platform attribution.
• Python package can be made reproducible in a governed modelling workflow.

Limits

• No official MCP, plugin or dedicated CLI was verified.
• Requires statistical skill and suitable historical data.
• It is not a daily campaign-operation interface.

Check the sources

Supported actions

Prepare

Partial

Configure

Yes

Run

Yes

Validate

Yes

Scenarios

Yes

Produces MMM results and allocation hypotheses, without a governed connection to live accounts.

Why this grade

• Established open-source marketing-mix modelling project from Meta.
• Supports automated model exploration and budget-allocation analysis.
• R and Python routes broaden analyst access.

Limits

• No official MCP or marketer-facing CLI was verified.
• Model selection and calibration still require specialist judgement.
• Outputs do not translate directly into safe platform changes.

Check the sources

Supported actions

Prepare

Partial

Configure

Yes

Run

Yes

Validate

Yes

Scenarios

Yes

Flexible modelling toolkit, not a campaign-operation or turn-key daily marketer interface.

Why this grade

• Flexible Bayesian tooling for media-mix and customer-value modelling.
• Open source and inspectable for teams that need methodological control.
• Python ecosystem supports custom analysis and reproducibility.

Limits

• Requires strong Python and statistical expertise.
• No official MCP, plugin or dedicated workflow CLI was verified.
• It is a modelling library rather than a daily media-operations product.

Check the sources