Genna
MCP MCP Connector Terms Terms of Service Privacy Privacy Policy

Genna Documentation

Genna MCP Connector

Last updated August 10, 2026 Effective August 10, 2026 Version 1.0

On this page

  1. What is the Genna connector
  2. How it works: you stay in control
  3. Connect your AI assistant to Genna
  4. Permissions & security
  5. What you can do
  6. Design rendering in the design workflow
  7. Publishing safeguards
  8. Data access & privacy
  9. Usage limits & credits
  10. Example prompts
  11. Troubleshooting
  12. Support & contact
On this page
  1. What is the Genna connector
  2. How it works: you stay in control
  3. Connect your AI assistant to Genna
  4. Permissions & security
  5. What you can do
  6. Design rendering in the design workflow
  7. Publishing safeguards
  8. Data access & privacy
  9. Usage limits & credits
  10. Example prompts
  11. Troubleshooting
  12. Support & contact

What is the Genna connector

The Genna connector lets you manage your brands’ social media content directly from your AI assistant, using the Model Context Protocol (MCP). With it, your assistant can plan, draft, and design posts for your brands, organize your media gallery, research reference accounts, read your own pages’ performance analytics, and — only when you say so — publish or schedule content to your connected social channels.

It is a remote MCP server operated by Genna Technologies, Inc. at https://mcp.genna.co/mcp, using the Streamable HTTP transport and OAuth 2.1 authentication. It works with any MCP-compatible AI assistant or agent that supports remote connectors, in chat apps and developer tools alike.

Using the connector requires a Genna account. Everything the connector does is governed by the Genna Terms of Service and the Genna Privacy Policy.

How it works: you stay in control

Genna is an assistive tool, and the connector follows the same principles as the Genna app:

  • Drafts first. Creating a post produces a draft inside your Genna workspace — nothing is posted anywhere. You can review every draft in the chat or open it in the Genna editor.
  • Human review and approval. Content reaches a social network only when you explicitly ask your assistant to publish or schedule it. The connector marks publishing tools as destructive so compliant assistants ask for your confirmation before calling them.
  • Scoped permissions. Publishing requires a separate mcp:publish permission that you grant (or withhold) when you connect. Without it, the connector can create and edit drafts but can never post.
  • Your accounts only. The connector acts only on brands in your Genna workspace and social accounts you have connected and are authorized to manage.

Connect your AI assistant to Genna

Chat assistants (custom connectors)

  1. Open your assistant’s connectors or integrations settings and choose to add a custom/remote MCP connector.
  2. Enter the endpoint: https://mcp.genna.co/mcp.
  3. You are redirected to Genna’s secure sign-in and consent screen. Sign in, choose which brands the connector may access, and which permissions to grant (read, write, publish).
  4. Approve. Your assistant is now connected — start a chat and ask it about your brands.

CLI and developer MCP clients

Add the connector with your client’s standard flow, using the endpoint https://mcp.genna.co/mcp and the Streamable HTTP transport. The OAuth flow starts automatically on first use.

Advanced users can alternatively authenticate with a Genna API key created in the Genna app, sent as a Bearer token in the Authorization header.

Disconnecting

You can revoke the connector’s access at any time: remove the connector in your assistant’s settings, or revoke the grant from your Genna account settings. Access grants expire automatically after 12 months, and idle refresh tokens after 90 days.

Permissions & security

  • OAuth 2.1 with authorization-code flow and mandatory PKCE (S256), served over TLS with certificates from recognized authorities.
  • Granular scopes: mcp:read (view brands, posts, analytics), mcp:write (create and edit drafts, upload media), mcp:publish (publish or schedule to connected channels). Unknown or empty scopes are rejected.
  • Brand allowlist: each grant or API key can be restricted to specific brands; ownership is re-verified on every call.
  • Short-lived tokens: access tokens expire after 30 minutes; refresh tokens rotate on every use, and reuse of an old token revokes the entire grant.
  • Credentials hashed at rest: API keys and refresh tokens are stored only as SHA-256 hashes.
  • Full audit trail: every tool call is recorded in your account’s activity history — identifiers only, never your conversation content.

What you can do

CapabilityWhat the connector does
Create posts (AI)Draft posts from an idea, a link, a product, an image, trending news, or a reference post — in your brand’s formats, tones, and hooks. Drafts are text-first; you review before any visuals are produced.
Create posts (manual)Turn media you already have (photos, videos) into a post or carousel with a caption, ready to schedule.
Design & refine visualsRender the slide images of a post design, then refine individual slides: adjust with an instruction, re-roll, or swap in a gallery image.
Manage the galleryUpload images and videos to your brand’s media gallery and browse albums.
Publish & scheduleWith the publish permission and your confirmation: publish now or schedule for later on your connected channels.
Social connectionsConnect or disconnect social accounts (Instagram, Facebook, and more) via each platform’s official authorization flow.
AnalyticsRead your own connected pages’ performance: per-post reach, likes, saves, shares, views, interaction rates, and daily time series.
ResearchKeep a reference library of public accounts for inspiration, validate Instagram post URLs, and browse industry news radar.
Content formatsTeach Genna new post formats from example posts and manage your brand’s format library.

Design rendering in the design workflow

The Genna connector is design-focused software: it produces social post designs — slides with your brand’s identity, typography, and layout:

  • There is no standalone image generator. No tool takes a free-form prompt and returns an image.
  • Images are only ever rendered for the slides of an existing post draft — the post’s title, caption, per-slide texts, format, and brand identity drive the visuals.
  • Per-slide refinements (adjust, regenerate, replace) operate on one slide of one post and are bounded by revision limits.
  • Uploading media to the gallery ingests files you already have; it generates nothing.

In short

Genna generates images the way a slide-design tool does: as part of producing a designed asset you asked for — never as a general-purpose image generation service.

Publishing safeguards

  • Publishing and scheduling require the separate mcp:publish permission.
  • Both tools are annotated as destructive, so compliant assistants confirm with you before calling them.
  • Publishing fails safely if the brand has no connected social account.
  • Every publish is additionally recorded as a dedicated audit event in your account history.
  • Scheduled posts can be reviewed and managed in the Genna app before they go out.

Data access & privacy

The connector accesses only the data needed to perform the action you asked for, within the brands you allowed:

  • Accessed on request: your brand profiles, posts and drafts, media gallery, connected social accounts, your own pages’ analytics, your products, audience personas, saved reference accounts, and industry news items.
  • Never accessed: your conversations with your assistant, its memory, or uploaded files beyond the tool parameters the assistant sends; other users’ workspaces; private content of any third-party account.
  • Logging: tool calls are audited with identifiers only — no conversation data and no generated content is stored in audit logs.
  • Media uploads go directly to Genna’s cloud storage over pre-signed URLs; server-side URL fetches are validated to block private networks and unsafe schemes.

How Genna collects, uses, and retains personal information is described in the Genna Privacy Policy. For data-protection questions, contact privacy@genna.co.

Usage limits & credits

  • Requests are rate-limited per credential to keep the service healthy.
  • Rendering a post’s images debits one generation credit from your Genna plan, exactly as in the Genna app. The assistant is informed of this in the tool description, and the connector fails with a clear message if your plan has no credits left.
  • Per-slide AI refinements draw on each slide’s revision allowance; swapping in a gallery image is always free.
  • Responses are kept compact by design; media previews are size-capped, and you can request a single slide instead of a full post to save tokens.

Example prompts

  • “Create an Instagram carousel for my brand about our new product launch — suggest a few ideas first, then draft it and show me the images before anything is published.”
  • “Here are three photos from today’s event — post them as a carousel on my Instagram with a caption you write, and schedule it for tomorrow at 9am.”
  • “How did my last 10 posts perform? Which format got the best save rate, and what should I post next based on that?”
  • “Make slide 2’s image warmer and remove the text overlay, then publish the post.”

Troubleshooting

IssueWhat to do
Connection failsVerify the endpoint is exactly https://mcp.genna.co/mcp and retry the OAuth flow. Remove and re-add the connector if the consent screen does not appear.
”Unauthorized” errorsYour access grant may have been revoked or expired. Reconnect from your assistant’s connector settings. For API keys, confirm the key is active in the Genna app and sent as Authorization: Bearer.
Publish is refusedPublishing requires the mcp:publish permission — reconnect and grant it — and at least one social account connected to the brand in Genna.
”Not enough credits”Rendering post images requires generation credits on your Genna plan. Top up or upgrade in the Genna app, then retry.
Brand not foundThe credential’s brand allowlist may exclude that brand. Reconnect and select the brands you want your assistant to access.
Long-running stepsPost creation and image rendering are asynchronous: the assistant receives a job id and polls it. If a wait times out, ask it to check the job again — the work continues server-side.

Support & contact

Product & security support — support@genna.co

Privacy — privacy@genna.co

Entity — Genna Technologies, Inc., a Delaware corporation · 131 Continental Dr, Suite 305, Newark, DE 19713, United States

See also: Terms of Service · Privacy Policy

FAQ Sitemap Terms of Service Privacy PT

© 2026 Genna AI. All rights reserved.