Schedule social media posts with n8n
Connect an n8n workflow to Bolta with an API key. Add the MCP Client Tool node to an AI Agent so it can draft posts in your brand voice, or use an HTTP Request node to call the Bolta REST API with fixed steps. Nothing goes out until a person approves it.
Nothing is approved, scheduled or published until you say yes.
Checked 2026-11-21
Create an API key in Bolta, save it as a Bearer credential in n8n, then run one test item.
Draft a LinkedIn post from this new blog item, in my brand voice, and save it as a draft for my review
Illustration of the reply: Bolta returns a draft showing the connected account and the scheduled time, with Approve, Edit and Reject actions.
How an n8n workflow uses Bolta
An n8n workflow can reach Bolta two ways. On the agent path, an AI Agent node picks Bolta tools itself. On the fixed path, an HTTP Request node makes the same call each run. Either way the result is a draft waiting in Bolta.
- 01
A trigger starts the run
A schedule, an RSS item, a form answer or a webhook starts the workflow in n8n. Whatever text it carries becomes the raw material for the post.
- 02
The workflow finds the brand
The agent lists the brands the key can reach, then the accounts connected to the one it picks. With more than one brand it needs a workspace_id, from your System Message or the workflow input.
list-workspaces - 03
It reads the posting rules
Before writing, the agent can check the limits that apply to the account it is drafting for, so a long text does not fail later.
get-posting-rules - 04
It learns the voice
The agent pulls the brand voice context so the draft sounds like the brand and not like a generic model.
get-voice-context - 05
It saves a draft
The text is saved as a draft on the chosen account. A draft is the end of the unattended part.
save-post - 06
A person decides
You open Bolta, read the draft and approve it, change it or reject it. Nothing is approved, scheduled or published until a person says yes.
list-inbox-items
Set up n8n
Connect once. Everything below works from n8n afterwards.
- 01
Create a Bolta API key
In Bolta, open Settings, API Keys and create a key for the brand this workflow should post for. For drafts and reading, tick Read Workspace, Read Accounts, Read Posts and Write Posts. The key starts with bolta_sk_ and is shown once, so copy it now.
- 02
Save the key in n8n as a credential
In n8n, open Credentials and create a Bearer credential. Paste your bolta_sk_ key as the token and name it something clear. If you use a Generic header instead, the name is Authorization and the value is Bearer followed by your key.
- 03
Path A: add the MCP Client Tool to an AI Agent
Click the plus under the Tools connector of your AI Agent node and add MCP Client Tool. Set the endpoint to https://mcp.bolta.ai/mcp, choose HTTP Streamable if the node offers a transport, set authentication to Bearer and pick your credential. HTTP Streamable arrived in n8n 1.104, so update n8n if you only see SSE.
- 04
Path A: tell the agent to save drafts only
In the AI Agent node, add a System Message that says to save drafts only and never approve, schedule or publish unless a person has said yes. Ask it to name the brand, or pass the workspace_id you give it, when more than one brand is available.
- 05
Path B: or call the REST API with an HTTP Request node
For the same call every time, add an HTTP Request node with Bearer or Header Auth and your credential. Pick the endpoint from the Bolta API docs. Run a read-only call first to confirm the key works, then run the workflow on one test item and find the draft in Bolta before you switch it on.
Connect n8n to Bolta
Create an API key in Bolta, save it as a Bearer credential in n8n, then run one test item.
Which Bolta connection methods work with n8n
n8n offers two documented routes, and this table separates what we have written up from what we have not confirmed.
| Method | With n8n | How |
|---|---|---|
| MCP with an API key | Works | Add the MCP Client Tool to an AI Agent, use https://mcp.bolta.ai/mcp, set authentication to Bearer and pick a credential holding your bolta_sk_ key. Needs n8n 1.104 or newer for HTTP Streamable. |
| REST API | Works | An HTTP Request node with Bearer or Header Auth calls the Bolta REST API. Each endpoint and its permission is listed in the Bolta API docs. |
| Hosted MCP with sign-in | Not confirmed | n8n's own page lists OAuth2 as a way to sign in, but we have not written up or tested a sign-in route to Bolta, so use the API key. |
| Skill via npx skills add | Not available | n8n is a workflow builder, not a coding agent, so there is no place to install a skill. |
| Plugin or extension | Not available | We have no Bolta plugin for n8n. The MCP Client Tool node or an HTTP Request node does the same job. |
What you can run in Bolta from n8n
Each row is an area of Bolta, with the real tools behind it.
Draft posts from any workflow event
A new RSS item, form answer or customer win can become a draft on a connected account. The draft lands in Bolta, not on the network.
save-postbulk-create-postsWrite in the brand voice
An AI Agent can read the brand voice before writing, so drafts from automation match drafts from people. Without that read it writes generically.
get-voice-contextvoice-generateCheck the rules first
The workflow can read what an account allows before it writes, which avoids drafts that are too long or the wrong kind for that network.
get-posting-ruleslist-accountsReport on what is waiting
A scheduled workflow can pull the approval queue and post a note to your team in chat or email, so drafts do not sit unseen.
list-inbox-itemslist-workspace-postsPull results into a sheet
Recent post performance can flow into a spreadsheet or report on a schedule, as long as the key carries the read permission for it. A 403 means one is missing.
get-account-analyticsget-postRun a routine from a trigger
If you already have a Bolta routine, a workflow can start it on an event. Routines have their own review rules, and a routine only runs after you confirm its scope in Bolta.
list-routinesrun-routine-now
A sample AI Agent run in n8n
An illustration of an AI Agent node with the MCP Client Tool attached, handling a new blog item. The agent stops at a draft.
- You
New blog item: 'Five checks before you hit publish'. Draft a LinkedIn post from it for the Harbour Bakery brand. Save a draft only.
- Tool call
list-workspacesFinds the brand the key reaches
- n8n
I found the Harbour Bakery brand. I will read its LinkedIn account and voice before writing.
- Tool call
list-accountsLooks for a connected LinkedIn account
- Tool call
get-voice-contextReads the brand voice
- Tool call
get-posting-rulesReads the LinkedIn limits
- n8n
Here is the draft: 'Before we publish anything, we run five quick checks. Number three catches most of our mistakes...' It follows the voice notes and fits the LinkedIn limit.
- Tool call
save-postSaves the text as a draft
- n8n
Saved as a draft in Bolta. Nothing is approved, scheduled or published. Open Bolta to review it, and tell me if you want a different angle.
- You
Looks right, I will approve it in Bolta myself.
What can n8n do with Bolta?
- Turn a new RSS item or blog post into a draft for LinkedIn or X
- Send a form answer or customer win to Bolta as a draft for review
- Let an AI Agent draft a week of posts in your brand voice
- Pull recent post results into a sheet or report on a schedule
- Check what is waiting for approval and post a note to your team
What is the Bolta MCP?
The server n8n talks to when it works with Bolta.
MCP, the Model Context Protocol, is the open standard that lets an AI app call tools in another product. Bolta runs a hosted MCP server at https://mcp.bolta.ai/mcp. Your AI app discovers its 67 tools and calls them for you: list your brands, save a draft, plan a routine, read results.
One connection is one organization. You tick the brands the app may reach, and every call is checked against your role, so the app can never do more in Bolta than you can. Approving, scheduling and publishing need your yes.
Brands & accounts6 tools
The brands this connection reaches, who works on them and who can approve, their approval settings, connected social accounts, and each platform's posting rules.
list-workspaces, list-workspace-members, set-brand-approval, list-accounts, get-connect-link, get-posting-rules
Posts9 tools
Save a draft or edit a post with save-post, then publish it now or at a time with publish-post after the user says yes.
list-workspace-posts, get-post, save-post, publish-post, delete-post, bulk-create-posts, create-wordpress-blog-post, list-recently-deleted, restore-deleted
Review4 tools
Everything waiting for a yes: submit drafts to the team or send them to a named person for approval, approve one or many, reject with a reason, and work the inbox of routine and campaign drafts.
list-inbox-items, submit-for-review, approve-many-posts, reject-post
Plans6 tools
Have Bolta plan a routine or a campaign the way its own chat does, change it, then save it on the user's yes.
plan, plan-routine, plan-campaign, recommend-routines, revise-plan, confirm-plan
Routines10 tools
Work that repeats and runs in Bolta even with the app closed.
list-routine-templates, list-routines, get-routine, create-routine, update-routine, activate-routine, pause-routine, archive-routine, run-routine-now, list-routine-runs
Campaigns6 tools
One to four weeks of posts about one thing: a launch, an event, a season.
list-content-campaigns, get-content-campaign, create-content-campaign, update-content-campaign, analyze-content-campaign-source, set-content-campaign-status
Team versions4 tools
One idea written as one post per person on the team, each in their own voice.
plan-team-posts, update-team-post-plan, generate-team-post-drafts, get-team-post-plan
Voice & business8 tools
The brand's voice and business facts, and writing in that voice.
get-voice-context, get-voice-profile, create-voice-profile, update-voice-profile, list-business-dna, extract-business-dna, update-business-dna, voice-generate
Suggestions2 tools
What Bolta suggests changing, for a routine or the whole brand, from what people approved, edited and rejected.
list-suggestions, decide-suggestion
Replies10 tools
Find conversations worth joining, draft replies in the brand voice, and send them once the user says yes.
search-engagement-opportunities, list-engagements, get-social-content, draft-engagement-reply, draft-inbound-reply, publish-engagement-reply, reply-to-social-content, react-to-social-content, send-outreach-reply, update-outreach-reply
Results2 tools
How posts and accounts are performing.
list-account-insights, get-account-analytics
Prompts for your AI Agent node
Put these in the agent's System Message or pass them in from the trigger. Each leaves a draft.
Use the item text from the trigger to draft one LinkedIn post in the brand voice. Save a draft only.
I read the voice and the LinkedIn limits, wrote one post and saved it as a draft. It is not approved or published.
get-voice-context, save-post
List the brands this key can reach and the accounts connected to each. Change nothing.
I found your brands and their connected accounts and left everything untouched. Tell me which brand to use.
list-workspaces, list-accounts
Show what is waiting for approval and write me a three-line note I can post to the team channel.
I listed the items waiting and drafted a short note for you. I did not approve anything.
list-inbox-items
Draft three X posts about this release note and save each one as a draft.
I saved three drafts on the X account, each under the post limit. None are approved or published.
get-posting-rules, save-post
Look at my last five published posts and tell me which one did best, in two sentences.
I read the recent results and named the strongest post with a short reason. I changed nothing.
list-workspace-posts, get-account-analytics
Pick the right brand from the form answer before drafting. If it is unclear, stop and ask.
The form did not name a brand, so I stopped and asked you to choose one instead of guessing.
list-workspaces
What keeps running when n8n is idle
n8n is the thing that runs unattended. Whether it keeps going depends on where it lives.
- A workflow that is switched on with a schedule or webhook trigger runs without you opening anything, as long as your n8n instance is running. We have not checked n8n Cloud limits, so check yours.
- Each run can save drafts to Bolta on its own, and that is where it stops.
- A Bolta routine you confirmed keeps running on Bolta's side, independent of n8n.
- The AI Agent path also needs the chat model you attached in n8n to be reachable on each run.
Unattended runs only save drafts. Approving, scheduling and publishing wait for a person in Bolta.
Agent path or fixed path in n8n
The two routes behave differently inside a workflow, and the difference shows up when something odd arrives.
The MCP Client Tool path lets the AI Agent choose which Bolta tools to call, so it can handle an item that needs a brand lookup first. The cost is that its steps vary from run to run.
The HTTP Request path makes one known call, so every run looks the same and is easy to debug. It cannot adapt on its own when the input changes.
Keep the key in a credential
- Create the Bearer credential once and reuse it in every Bolta node.
- Tick only Read Workspace, Read Accounts, Read Posts and Write Posts for drafting.
- Use one key per brand when the workflows are for different clients.
- Revoke and replace the key in Bolta if it appears in a shared export.
Test on one item first
- 01
Run a read-only call
List workspaces to confirm the key works and to find your brand's workspace_id.
- 02
Execute on one test item
Click Execute workflow once and open Bolta to find the draft.
- 03
Switch it on
Turn the workflow on only after the draft looks right.
What n8n cannot do without you
Drafts are free, finals need your yes
n8n can write, edit and save drafts at any time. Approving, scheduling, publishing, sending a reply, activating a routine or campaign and deleting each wait for your yes, shown against the exact text.
One connection, one organization
When you connect n8n you tick the brands it may reach. It reaches only those, and when that is more than one brand, every call names the brand it is for.
Your role is the ceiling
Each call n8n makes is checked like a normal request from you. Owners and admins can act on any post, creators on Bolta's drafts and on posts they are a named reviewer on, and viewers never approve or publish.
Where n8n can publish through Bolta
11 social networks, plus WordPress for articles. Bolta holds the connected accounts. Each card opens the n8n guide for that network.
- Threadstext posts, image posts, polls
- LinkedIntext posts, articles, carousels
- X (Twitter)tweets, threads, polls
- Instagramposts, stories, reels
- Facebookposts, stories, reels
- Blueskyposts, threads, quote posts
- Mastodontoots, polls, threads
- Reddittext posts, link posts, image posts
- TikTokvideos, stories, photo posts
- YouTubevideos, Shorts, community postsTeam plan, beta
- Pinterestpins, idea pins, video pinsTeam plan, beta
The words behind n8n and Bolta
- MCP
- The Model Context Protocol, an open standard that lets an AI app call tools in another product. Bolta's hosted server is at https://mcp.bolta.ai/mcp, and n8n reaches Bolta through a connection like it.
- Skill
- A folder of instructions an AI app reads so it knows how to use Bolta well. It does not connect by itself: install it with
npx skills add boltaai/bolta-skills --skill boltanext to the connection. - Plugin
- A bundle that brings the connection and the skill in one install, where the app supports plugins. Claude Code has one, added with its plugin marketplace commands.
- API key
- A
bolta_sk_key you create in Settings, sent as a Bearer header. It carries the permissions you choose, which suits apps you configure by hand or anything that runs unattended.
Setup for n8n is not marked as confirmed against official docs. Last checked 2026-11-21.
Other AI agents that post through Bolta
Every agent below connects to the same Bolta workspace and follows the same approval rules.
- ChatGPTOpen Bolta in ChatGPT, sign in
- OpenAI dotsAlways-on ChatGPT agents
- ClaudeCustom connector, one sign-in
- Claude DuetCarries your Claude connectors over
- Claude CoworkPlugin and connector on your desktop
- Claude CodeTwo plugin commands
- Codexcodex mcp add, then sign in
- CursorOne mcp.json entry
- GeminiGemini CLI extension
- OpenClawSkills and MCP for your gateway
- Hermes AgentMCP server in the Hermes config
- Grok BotCustom connector in xAI's agent
- Grok Buildgrok mcp add
- DeepSeekThrough the Bolta MCP link
- Kimi K3Kimi Code with the Bolta MCP server
- MuseMeta's agent, Muse Code and connectors
- ManusConnector with an API key
- CueA personal agent that posts for you
- PerplexityResearch there, draft in Bolta
- Perplexity ComputerRemote MCP connector
- nanoclawRun it from WhatsApp or Telegram
- PaperclipConnector for your agent company
- MCP ServerOne link for any MCP client
What n8n can do through Bolta
The jobs Bolta's agents cover, each with its own page. Your AI app can ask for drafts, plans and reports in the same areas. Every post still waits for your yes.
Frequently asked questions
Short answers about using n8n with Bolta.
- Does Bolta publish without my approval?
- No. A workflow can save drafts on its own, but nothing is approved, scheduled or published until a person says yes. Approving stays with you in Bolta, and Safe Mode and client sign-off still apply.
- Should I use the MCP Client Tool or the HTTP Request node?
- Use the MCP Client Tool if you want an AI Agent to decide which Bolta tools to use. Use the HTTP Request node if you want the same fixed call every time. The MCP path needs n8n 1.104 or newer. If your node has no HTTP Streamable choice, use the HTTP Request node.
- Is this tested on every n8n version?
- No. Menu and field names change between n8n versions, so look for the closest label if one does not match. The AI Agent path also needs a chat model you set up in n8n yourself, which is separate from Bolta.
- What can it not do?
- A key can only do what the permissions you ticked allow, and a call without a permission returns a 403. Posting to a network needs that account connected in Bolta first. If the connection reaches more than one brand, each call has to name the workspace_id.
- How do I keep the API key safe?
- Treat it like a password. Keep it in an n8n credential, not in a Set node or in the workflow text. If it leaks, revoke it in Bolta and make a new one.
- What does the API key start with and where do I make it?
- Bolta keys start with bolta_sk_. You create one in Bolta under Settings, API Keys, and it is shown once, so copy it straight into an n8n credential. If you lose it, revoke it and make another.
- Which permissions should the key have?
- For drafting and reading, tick Read Workspace, Read Accounts, Read Posts and Write Posts. Leave the rest off. A call that needs a permission the key lacks returns a 403, and the fix is to edit the key or make a new one.
- What does the node answer if the key is wrong?
- The MCP node reports Invalid or expired API key, and an HTTP Request node returns a 401. Check the credential sends Authorization with the value Bearer bolta_sk_ followed by your key, with one space after the word Bearer. Sending the same key again gives the same answer, so replace the key.
- Why does the agent ask which brand to use?
- A key can reach more than one brand, and in that case every call has to name the brand. The API calls a brand a workspace, so give the agent the workspace_id in its System Message, or tell it the brand name in the workflow input.
- The agent cannot see any Bolta tools. What is wrong?
- Check Tools to Include on the MCP Client Tool node. If it is set to Selected with nothing chosen, nothing is exposed, so set it to All or pick the tools you want. Also confirm the node is attached to the Tools connector of the AI Agent.
- Can a workflow approve or schedule posts by itself?
- It can save drafts without anyone watching. Approving and scheduling stay with a person, and the System Message we suggest tells the agent to stop at a draft. Treat that message as a guard rail and rely on the Bolta approval rules for the real limit.
- Can one workflow serve several clients?
- Yes, if the key reaches several brands and each call names its workspace_id. A tidier option is one key per brand, saved as separate n8n credentials, so a mistake in one workflow cannot touch another brand.
Ready to schedule social media posts with n8n?
Connect n8n and ask for your first draft. Nothing goes out until you say yes.