One API for LinkedIn content inside your AI agent.
Connect Claude, ChatGPT, Cursor, or any MCP-compatible agent to Postiv. Search workspace knowledge, draft posts, generate images and carousels, schedule with approvals, and read analytics, all through one server.
Works with
$ claude
Pulled positioning, past posts, and brand rules from the workspace.
Draft ready, grounded in real workspace context.
Narrative, slide structure, and copy ready to render.
Queued for publishing and routed to any configured approvers.
All of Postiv, exposed as LinkedIn content tools.
The MCP turns product workflows into agent-callable tools, so your agent can create useful LinkedIn work without fragile prompts or a chain of disconnected services.
Workspace context
get_workspace confirms the connected workspace and which scopes the key can use before an agent does anything else.
Knowledge grounding
add_knowledge and search_knowledge let an agent save facts and pull real positioning instead of guessing.
Inspiration research
search_inspiration searches a moderated library of what is actually performing on LinkedIn before a first draft.
Profiles and Comment Plug
list_linkedin_profiles, get_comment_plug_config, and set_comment_plug_config manage which profile to post as and its automatic first comment.
Post drafting and lifecycle
create_post, list_posts, get_post, and update_post cover the full draft lifecycle for an agent.
Scheduling with approvals
schedule_post, reschedule_post, and unschedule_post queue posts for publishing and respect any human approvers on the target profile.
AI image generation
generate_post_image and edit_post_image create a branded image for a post, then iterate on it with plain-language instructions.
Carousel generation
create_carousel, render_carousel, and edit_carousel_slide turn a structured outline into a finished, branded LinkedIn carousel.
Performance analytics
list_analytics_accounts, get_analytics_summary, get_performance_overview, get_top_posts, and get_post_analytics read cached LinkedIn performance data.
Interactive widgets
In hosts that support MCP Apps, such as claude.ai, tools render inline widgets: a post editor, image preview, carousel pager, and inspiration cards.
LinkedIn-safe formatting
LinkedIn renders plain text. Bold and italic markers are converted to LinkedIn-safe formatting automatically, so posts display the way they were written.
Nothing publishes by surprise
Posts never go live instantly. schedule_post always queues the post, and it waits for approval if the target profile has approvers configured.
Connect Postiv to your AI agent
Get an API key from your Postiv workspace, then point your agent at the MCP server. Same key, same workspace, whether you use claude.ai, Claude Code, or another MCP client.
Step 0. Get your API key
In Postiv, go to Settings, then Integrations, then the Public API section, and generate a key. The same key works for the MCP server and the REST API. Only admins can generate or view it.
Get your API key
Add Postiv as a custom connector in claude.ai
claude.ai custom connectors can only take a URL, they cannot send a header, so the API key lives in the URL itself. Treat this URL like a password.
- 1. In claude.ai, go to Settings, then Connectors.
- 2. Click "Add custom connector".
- 3. Paste your Postiv URL with your key in it.
https://postiv.ai/mcp/pk_postiv_YOUR_KEY
Full endpoint reference, scopes, and the REST API live in the API documentation.