Connect Claude to your marketing data from Google and Meta, enabling read and write operations on Search Console, Analytics, Tag Manager, Business Profile, and Meta platforms.
Provides seven inspectable marketing workflows covering ad tagging, survival summaries, brief scoring, search-term classification, fatigue detection, landing-page promise matching, and lead fit scoring via a stdio MCP interface.
Enables AI apps to plan, build, audit, and report on Google Ads, GA4, Search Console, and Tag Manager accounts through natural language. Accounts are read-only by default, and every change is validated in a preview, applied only after confirmation, logged, and reversible.
Enables marketing optimization tasks such as copywriting, campaign analysis, social media planning, audience segmentation, and KPI tracking through natural language.
Enforces a marketing pre-flight: plan the campaign with kill criteria, score each offer on an explainable 0-100 rubric, and require every tactic to cite a scored offer plus a success metric before any spend.
Enables Pinterest Business account management through OAuth, including authorization URL generation, connection status checking, and secure token storage.
Connect any AI agent to your personal Telegram messages through the official Business API, enabling message history search and draft replies with optional manual approval.
Enables developers and product owners to scan a codebase and extract a readable business overview—features, roles, access control, data model, and endpoints—along with findings on missing or inconsistent rules, and generate shareable HTML or Markdown reports and diagrams.
Exposes the OSMOS Marketing API registry as callable MCP tools, enabling AI agents to discover endpoints and make real API calls with authentication headers and campaign hierarchy support.
Unifies Google Ads, Tag Manager, Search Console, and GA4 into a single MCP server, enabling reporting, monitoring, campaign/tag/property creation, and live mutations with dry-run safety.
Enables AI agents in Codex or Claude Code to research French companies through tools for company searches, mandates, BODACC announcements, INPI acts and accounts, published decisions, and relationship graphs, with a method to distinguish facts, homonyms, and leads to verify.
Enables managing a Yandex Business branch network through an XML feed from any MCP client. Agents can edit branch data in YAML, import/export, validate, diff, and build the feed locally, while publishing to Yandex remains manual.
Enables querying and analyzing marketing data from Google Analytics 4, Search Console, YouTube, and optionally Google Ads and Meta Ads, stored in a unified DuckDB database, via natural language in Claude.
Runs a local stdio MCP server that lets you evaluate an AI agent's business outcome against four synthetic scenarios, including a duplicate-free quote follow-up, an unsent invoice reminder, client-file separation, and a stockout. It scores the run from the final state and recorded attempts rather than the agent's text, returning a pass or fail exit code.
Provides read-only tools to search customers, orders, and business documents across PostgreSQL, JSON CRM, and Markdown sources, with tenant isolation, PII masking, and auditable access.
An MCP server that lets agents collect a company's entire public marketing footprint from a single URL — paid ads, organic social, SEO, AI-crawler stance, Reddit and founder activity, tracking pixels, and funnel signals — and return a compact scorecard plus an evidence index with unknowns. It also exposes tools to pull the underlying ad copy, posts and media, retrieve the analysis playbook, diff two runs over time, and validate that a generated report's citations resolve without invented spend.
Enables AI assistants to analyze any website URL and return structured business data — name, logo, brand colors, merchant category code, headquarters address, phone, email and social links — each with a status, confidence, source and supporting evidence. Also exposes tools to explain step-by-step how a single field was decided, look up MCC codes, and report model usage.
