Announce what you ship
Claude Code already knows what changed. Ask it to turn the changes into posts without leaving the terminal.
Claude Code + Postcaly
Ship a feature, then ask Claude Code to announce it. Add Postcaly with one command, sign in once, and Claude Code drafts and schedules the posts in your Postcaly calendar.
Available now on every plan
Connects by Sign-in link (MCP)
Use Postcaly today: 14 days free, then from £15.99 a month. Updated
Look at the commits since the last tag and draft a release post for X and Bluesky. Save it as a draft.
There are 14 commits since v2.3.0; the main change is faster image uploads. I've saved one draft in Postcaly with a version for X (212 characters) and one for Bluesky (268 characters). Both passed Postcaly's checks. It's in your calendar as a draft.
Remote MCP server with sign-in
Claude Code adds remote MCP servers from the terminal. Postcaly is one of them.
Run the command below in your terminal. Or install the Postcaly plugin: /plugin marketplace add postcaly/skills, then /plugin install postcaly@postcaly.
Claude Code opens a Postcaly page in your browser. Pick one workspace and tick what it may do. Publishing straight away is off unless you tick it.
Ask Claude Code to write about what you just built. It checks each network's rules through Postcaly and saves drafts or schedules posts.
claude mcp add --transport http postcaly https://app.postcaly.com/mcpWhere there is no browser, Claude Code can send a Postcaly API key as a header instead. Create one in Postcaly under Settings, AI agents.
Plain requests work. Claude Code does the writing; Postcaly checks each network and keeps the schedule.
Write a short changelog post from this week's merged commits and draft it for LinkedIn and X.
We just tagged a release. Draft a Bluesky and Threads post announcing it, scheduled for 10:00 tomorrow.
Read docs/launch.md and plan three posts for the launch week, saved as drafts.
List what is scheduled in our Postcaly workspace for the next seven days.
Claude Code already knows what changed. Ask it to turn the changes into posts without leaving the terminal.
Signing in happens in your browser, so there is no token to paste into project files.
Plan limits, daily post limits and each network's checks apply to Claude Code exactly as they do to you.
The same actions as the Postcaly screens, with the same plan limits and network checks. You tick which permissions to give.
Looks, never changes
Drafts for a person
Books a stated time
Your library
Only if you allow it
Not included: editing or deleting posts that are already live, managing people or billing, and connecting social accounts. Those stay with a person in Postcaly.
Claude Code asks for access, you sign in to Postcaly and choose one workspace and what it may do. Each workspace is approved separately.
Posts are only created when you ask. A scheduled post needs a stated time, and publishing straight away is off unless you allow it.
Posts made by an agent carry the agent's name in your calendar, and every action is listed under Activity.
Switch on Drafts only for a workspace and everything an agent creates arrives as a draft for your team to check.
Your plan's limits and each network's rules apply to every agent, the same as in the Postcaly screens.
Remove the agent in Postcaly under Settings, AI agents, and its access stops at once. Removed team members' access stops too.
Claude Code already sits in your project, next to the release notes, commit history and docs. That is where a lot of product news starts. With Postcaly connected it can turn a release into posts without you copying text between windows.
It suits developers and small product teams who announce updates on X, LinkedIn, Bluesky or Threads and want the posts checked and scheduled from the terminal.
Claude Code is Anthropic's coding agent for the terminal and code editors. It connects to tools through MCP servers that you add with the claude mcp add command, and plugins can bundle MCP servers too.
Add Postcaly with one command, or install the Postcaly plugin from github.com/postcaly/skills, which adds the same connection plus a skill. Both use the same Postcaly account and do the same things.
A few habits make agent-made posts better and safer.
Some things are deliberately left out of the agent connection, because they cannot be undone or need someone's judgement:
Yes, on every Postcaly plan. Any account with a confirmed email address can connect it.
No. Claude Code signs in through your browser. If you prefer a key, for example on a server, create an API key in Postcaly under Settings, AI agents.
The channels connected in your workspace: today YouTube, X, LinkedIn profiles, Threads and Bluesky. Facebook Pages, Instagram, Google Business Profile, TikTok and Pinterest open as each network approves Postcaly.
No. Claude Code only creates posts when you ask it to. A scheduled post needs a time it states, everything lands in your Postcaly calendar where you can change or delete it, and publishing straight away stays off unless you allow it when you approve the connection.
Yes. Each person adds the server and signs in with their own Postcaly login, so Postcaly knows who made each post and applies their role in the workspace.
Claude Code can add an MCP server for a single project or for every project, using its own scope options. Choose whichever suits how your team works; the Postcaly sign-in is the same either way.
Nothing extra: agent access is included in every Postcaly plan. Your plan's limits on channels, posts and X posts apply to everything an agent does.
14 days free on every plan, then from £15.99 a month. Cancel any time.
Start your free trialWe'd like to use Google Analytics cookies to see how people use this website. Nothing is set unless you accept. Details