Automate social media posting with n8n
Attach the PostEverywhere MCP Client Tool to an AI Agent node and one workflow can write, schedule and publish to every account you have connected. No branch per network, no payload per platform to maintain.
$0 due today, cancel anytime
Used by 2,831 happy customers
New RSS item: “Release 3.2 adds team calendars.” Post it where it fits and schedule it for the next weekday morning.
Used PostEverywhere
list_accounts4 accounts foundget_platform_rulesNo image: skip Instagramcreate_post×3LinkedIn, X, Threads
Scheduled for Tuesday at 8:30 AM on LinkedIn, X and Threads. Instagram was skipped because the item had no image.
Mon
8:30 AM
Tue
8:30 AM
Wed
1:00 PM
Setup in n8n
Connect n8n to social media with the MCP Client Tool
n8n speaks the Model Context Protocol through its built-in MCP Client Tool, a sub-node that hands an AI Agent every tool on a remote server. Point it at PostEverywhere and the agent can see all 38 publishing tools, with the descriptions and schemas it needs to choose between them. Setup takes about two minutes.
- 1
Add an AI Agent node.
Create a workflow with a trigger (RSS Feed Trigger, Schedule Trigger, Form Trigger or a chat message) and add an AI Agent node with a chat model attached. The agent is what reads the input and decides which tools to call.
- 2
Attach the MCP Client Tool.
On the agent’s Tool input, click the plus, search for MCP Client Tool and add it. It is built in on n8n Cloud and self-hosted, so there is nothing to install from the community nodes. n8n’s MCP Client Tool docs list every parameter.
- 3
Paste the PostEverywhere endpoint.
Put the address below in the endpoint field. If your n8n version asks for a server transport, choose HTTP Streamable.
https://mcp.posteverywhere.ai - 4
Set Bearer authentication.
Choose Bearer Auth and create a credential holding a PostEverywhere API key from the Developers page (it starts with pe_live_). The node also offers header auth and OAuth2, but a revocable key suits workflows that run unattended better than a browser sign-in.
- 5
Run a test.
Send the agent “list my connected social accounts”. It calls list_accounts and returns each linked account with its status. If they come back, the connection is good and the agent can start scheduling.
What you need for n8n social media automation
Any n8n instance (Cloud or self-hosted) with the AI Agent node and a chat model credential, a PostEverywhere plan or active 7-day trial, at least one connected social account, and an API key with the write scope. Add the ai scope if the workflow should generate images or captions.
Three workflows to build first
Three n8n social media workflows to build first
Each of these is a trigger, an AI Agent with the PostEverywhere tool, and a system message telling the agent what good looks like. The text the agent receives is shown as the request.
“RSS to social: “Here is a new blog post. Write a version for each connected account and spread them over the next three weekdays.””
RSS Feed Trigger into the agent. Replaces a fan of HTTP Request branches with one tool. Most people build this one first.
“Form to drafts: “Turn this product update into three posts and save them as drafts. Don’t schedule.””
Form Trigger into the agent, create_post with draft set to true. A person reviews and schedules the drafts in the calendar.
“Failure watch: “Which posts failed since yesterday? Retry each once and tell me which still fail.””
Schedule Trigger every morning, then retry_failed_posts. Send the leftovers to Slack or email with a normal n8n node.
“Weekly gap check: “Look at next week’s queue and fill any empty weekday with a post from this list of evergreen tips.””
Reads get_queue and list_posts, then schedules only where there is a gap.
Skills for n8n
Use the PostEverywhere skills as n8n agent instructions
n8n workflows do not install skills, but the AI Agent node takes a system message, and that is where a skill’s steps belong. Copy a skill from GitHub into the system message and the agent runs the same playbook every time the workflow fires, with the MCP Client Tool doing the posting.
Use the prompts as written
Or put one of the prompts below in the AI Agent node’s user message, with the details filled in by earlier nodes such as an RSS item or a form submission.
Repurpose content
Turn a blog post, transcript or newsletter into a set of posts.
Schedule social posts
Write, schedule or publish a post to any connected account.
Fix failed posts
Find posts that didn't go out, explain why, and fix them.
Weekly report
A short weekly or monthly report, ready to paste into an email.
Approval workflow
Draft posts for a client or manager to review before they go live.
Plan a content calendar
Plan a week or month of posts and fill your queue.
Prompts to try in n8n
See all skills and prompts“Turn this RSS item into a LinkedIn post and an X post with a link back to it, and schedule both in PostEverywhere for the next weekday at 9am.”
Skill: Repurpose content“Check PostEverywhere for posts that failed in the last 24 hours, retry the ones that can go now, and return a list of the rest.”
Skill: Fix failed posts“Save this form submission as a draft post in PostEverywhere for the brand team to review, with the headline as the first line.”
Skill: Approval workflow
What n8n schedules
One PostEverywhere node instead of eleven HTTP Request nodes
The usual n8n social workflow is a fan of branches, each with its own credentials, payload shape, media path and error route. It works for two networks that never change, and it degrades the moment you add a third or a platform revises its API. With PostEverywhere, the agent makes one kind of call for every network, and aspect ratios, caption limits and video rules are checked on our side. Everything the workflow schedules shows up in the calendar, where your team can see and edit it.
- Platform rules handled.
- get_platform_rules tells the agent each network’s limits before it writes, so a long LinkedIn post does not get forced onto X.
- Account health first.
- get_account_health catches an expired connection before a scheduled post fails, which n8n branches usually find out about too late.
- Signed webhooks back into n8n.
- Point a PostEverywhere webhook at an n8n Webhook node to start a workflow on post.published, post.failed or account.reconnect_needed.
Mon
12
8:30 AM
Blog: pricing explained
Tue
13
8:30 AM
Release 3.2: team calendars
Wed
14
1:00 PM
Customer photo of the week
Post to LinkedIn, Instagram and every top platform from n8n
The same tool call works for every account you connect in PostEverywhere. n8n never stores a LinkedIn or Instagram token, and you never maintain a separate node or app review per network.
Instagram
Instagram Graph API
Facebook
Meta Graph API
X
X API v2
LinkedIn
LinkedIn Marketing API
TikTok
Content Posting API
Threads
Meta Threads API
Bluesky
AT Protocol
Telegram
Telegram Bot API
Discord
Discord webhooks
YouTube
YouTube Data API v3
Pinterest
Pinterest API
WordPress
WordPress REST API
n8n limits and troubleshooting for unattended workflows
By platform
- Instagram64 sent
- LinkedIn48 sent
- X40 sent
- Facebook31 sent
- TikTok1 of 22 failed
- YouTube12 sent
Why posts failed
Grouped by cause, most common first
This video is outside the length TikTok allows for this account.
Example data
Agent node or plain API calls? Use the agent when the decision varies: which networks suit this item, how the caption should change, whether it is worth posting. When the logic is fixed (every post goes to the same four accounts), skip the model and call the social media API from an HTTP Request node: POST to app.posteverywhere.ai/api/v1/posts with content, account_ids and scheduled_for. Deterministic is cheaper and more predictable, and many teams run both.
Rate limits. Each API key gets 60 requests a minute and 1,000 an hour, on every plan. Bulk create takes up to 50 posts per call, but each post also counts, so a 50-post batch uses 51 of the minute’s 60. Space big batches with a Wait node, and see the rate limit guide.
Webhooks arrive at least once. PostEverywhere retries a delivery up to six attempts, so the same event can reach your Webhook node twice. Deduplicate on the X-PostEverywhere-Event-Id header and check the HMAC signature in X-PostEverywhere-Signature before acting. Verifying signatures has the code.
Unattended runs publish without a review. A chat window shows you each tool call; a scheduled workflow does not. While you test, tell the agent to save drafts or schedule for later rather than publish now. Scheduled posts can be edited or deleted until they go out; a published post usually cannot be recalled, because most platform APIs do not allow it.
The node will not connect. Check the endpoint has no trailing path, that the credential is Bearer with the full key, and that your PostEverywhere subscription is active. The server refuses calls from an inactive workspace instead of failing quietly.
The agent calls the wrong tool. Name the outcome in the system message (“schedule, never publish now”, “drafts only”). A large tool list rewards clear instructions. You can also limit which tools the node exposes if your version offers that option.
Why PostEverywhere in n8n
Agent node or plain API calls in n8n
POST /api/v1/posts
{ "content": "Launch week is here",
"account_ids": [101, 102, 103, 104, 105],
"scheduled_for": "2026-10-06T09:00:00Z" }- northwind.coffeedoneviews 1,204 · likes 87
- Northwind Coffeedoneimpressions 612
- northwindverifyingattempts 1
- northwindcoffeeuploadingattempts 1
- Northwind Coffeeretry_schedulednext_retry_at 09:05
One request, one result per accountExample data
PostEverywhere takes that layer out of the canvas. You connect accounts once in the app, PostEverywhere holds the tokens and publishes through each network’s official API, and n8n talks to one server. The decision about where and how to post moves from a tangle of IF nodes to an agent that reads the input and calls cross-posting tools with a reason for each choice.
For LinkedIn in particular, that means no LinkedIn app of your own to register and no token to refresh inside n8n. For Instagram, the agent learns from get_platform_rules that a post needs an image or video and skips it, or calls generate_image first, instead of sending a text-only post that would fail.
The other direction works too. n8n can publish your own workflows as an MCP server, so an assistant can call an n8n workflow as a single tool. That is a separate trigger node and a different job, covered in n8n’s guide to connecting to the n8n MCP server. Some teams combine the two: n8n handles enrichment and publishing through PostEverywhere, and exposes one high-level workflow to a chat assistant.
If you would rather not run a model at all, the same key works for the REST API, a typed Node SDK and the PostEverywhere CLI, and bulk scheduling in the app takes a CSV of up to 200 posts. The developer docs cover every endpoint, and building an agent on the API shows a full pipeline.
Every plan includes the API and the MCP server, from Lite at $9 a month. Start with the trial and have a working workflow the same afternoon.
$0 due today, cancel anytime
Connect another AI agent
Your n8n workflow and a chat assistant can share one PostEverywhere account and one calendar.
Pricing
Simple pricing, unlimited posts on every plan
Every plan posts to all top platforms and includes the API and hosted MCP server for your AI agent.
7-day free trial on every plan. Card required, $0 due today, cancel anytime. Compare every plan
FAQ
n8n social media automation FAQ
Straight answers about plans, platforms and how scheduling works.
Can n8n post to social media?
Yes. With the PostEverywhere MCP Client Tool on an AI Agent node, or with HTTP Request nodes calling the PostEverywhere API, n8n can schedule and publish to Instagram, LinkedIn, TikTok, YouTube, Facebook, X (Twitter), Threads, Pinterest, Bluesky, Telegram, Discord and WordPress through each network’s official API.
How do I use the MCP Client Tool in n8n?
Add it as a tool on an AI Agent node, set the endpoint to https://mcp.posteverywhere.ai and choose Bearer authentication with your PostEverywhere API key. It is a sub-node, so it only runs through an agent. For publishing with no model involved, use HTTP Request nodes against the REST API.
Can n8n post to LinkedIn and Instagram?
Yes, through PostEverywhere. Connect your LinkedIn profile or page and your Instagram professional account in the app, and the workflow can target them by account ID. Instagram posts need an image or video, which the agent can import with upload_media_from_url or create with generate_image. See the LinkedIn scheduler and Instagram scheduler for format details.
Do I need n8n Cloud or self-hosted?
Either. The MCP Client Tool is built into both, and the PostEverywhere server is reachable over HTTPS, so a self-hosted instance behind a firewall only needs outbound access.
How is PostEverywhere different from n8n’s own social nodes?
n8n’s per-network nodes each need their own developer app, credentials and maintenance, and they cover only some networks. PostEverywhere gives you one credential and one tool set for all top platforms, plus a calendar your team can see, health checks and retries.
Is there a free trial?
Every plan starts with a 7-day free trial. A card is required to start it, $0 is due today and you can cancel anytime. There is no free plan; API and MCP access are on every plan rather than sold as an add-on.
Still have questions?
Contact us or browse the full FAQ.
$0 due today, cancel anytime
Build your first n8n social workflow today
One node, one key, every account you connect.
$0 due today, cancel anytime
More for n8n workflows
Explore related features and guides to get the most out of PostEverywhere.
Social media API
The REST endpoints for HTTP Request nodes.
Webhooks
Signed events that start an n8n workflow.
Automate posting with the API
Worked automation examples.
Social media automation
What PostEverywhere automates for you.
AI agents
Every assistant and the full MCP tool list.
API vs MCP
When to call the API and when to use the MCP server.

“This tool has truly made my content posting so much easier, I can't work without it. I can schedule my posts days in advance while working on other projects. Set it and forget it.”