Abency
Run a marketing agency's social, SEO, ads, WordPress and email from an AI agent.
One tool, send_turn, hands a message to Enrich Labs' autonomous marketing agent.
Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.
Paste it into any MCP client. Setup docs
Helena is Enrich Labs' autonomous marketing agent, and this connector puts it inside Claude over OAuth. It exposes exactly one tool, send_turn, which passes your message to Helena and returns her reply. What she does in between is unbounded by that count: Enrich Labs says she publishes content, changes ad campaigns and runs sandboxed shell commands.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
initialize was refused with HTTP 401 and a Bearer challenge on 2026-09-08, which is the live check that established the OAuth posture. The name send_turn comes from Anthropic's directory listing. We have not seen its schema or its safety annotations, so this page makes no claim about either.send_turn call reaches.mcp. No scope names read separately from write, so no consent screen can offer a narrower grant and no issued token can be audited for one afterwards.Pro Plus, at $199 a month as listed on Enrich Labs' pricing page on 2026-09-08. The connector documentation says only that a paid plan is required, but the pricing page files "MCP access: use Helena from Claude" under Pro Plus alongside multi-brand management and Google Drive sync. Starter at $39 and Pro at $99 do not list it.
Enrich Labs documents no automatic gate. Its connector page says Helena can take real-world actions such as publishing content or changing ad campaigns when requested, and then instructs you to state constraints clearly and review drafts, budgets, dates and destinations before approval. That instruction is addressed to you, not to a system that blocks the action.
Because it is a conversational passthrough, not a toolset. The single tool hands your message to Helena, and Helena chooses from her own tools on the Enrich Labs side. Enrich Labs' engineering team published in July 2026 that Helena reaches roughly 24 core tools, a couple hundred native tools across about 50 platforms, and tens of thousands of actions across 3,000-plus apps.
It can reach the accounts that do. Enrich Labs states Helena launches, manages and optimizes campaigns on Google Ads, Meta and TikTok Ads when requested, using the ad accounts already connected to your Enrich Labs workspace. Enrich Labs publishes no spending cap, per-action limit or approval threshold for the connector in any documentation we could read.
Anthropic's directory entry says yes. Its permissions field for this connector states Helena can execute arbitrary shell commands in an isolated sandbox with outbound network access, and the description repeats the claim. Enrich Labs' own connector documentation does not mention shell execution at all, so the only source for it is Anthropic's listing.
No. Enrich Labs states that AI image, video and audio generation is not available through the Claude directory connector, and directs you to use Enrich Labs directly for media-generation workflows. Claude can still use the connector for text, analysis, planning and the other marketing actions. Enrich Labs charges 12 to 24 credits for video generation in its own product.
Helena reads the brand context and channels already connected inside your Enrich Labs account, so the connector does not add data sources on its own. Enrich Labs' privacy policy states your data is never used to train AI models. Its security page describes AES-256 encryption at rest and TLS 1.2 or higher in transit.
No. Enrich Labs uses OAuth 2.1 and states you never paste an API key into Claude. We confirmed the posture live on 2026-09-08: an anonymous request returned HTTP 401 with a Bearer challenge pointing at the server's own protected-resource descriptor, which names a single scope called mcp. You select a brand during authorization.
/integrations/slack, /integrations/notion, /integrations/google-drive (fetched 2026-09-08). · retrieved 2026-09-08initialize against https://agent.enrichlabs.ai/api/mcp — HTTP 401 with a Bearer challenge naming the protected-resource descriptor (2026-09-08). No tool list was read. · retrieved 2026-09-08https://agent.enrichlabs.ai/.well-known/oauth-protected-resource and /.well-known/oauth-authorization-server (fetched 2026-09-08). Single scope mcp; authorization code with PKCE. · retrieved 2026-09-08robots.txt on both hosts (fetched 2026-09-08): agent.enrichlabs.ai disallows only /marketing/ and /public/share-chat/; www.enrichlabs.ai allows everything except /demo. Neither carries a Content-Signal line. · retrieved 2026-09-08Connect Helena by Enrich Labs once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.