Typefully: Social Media Infrastructure for AI Agents

Typefully is a writing and scheduling tool for social media, supporting X (Twitter), LinkedIn, Threads, Bluesky, and Mastodon from a single dashboard.

What makes it interesting for developers: Typefully exposes a full API and an MCP server, so AI agents can draft, schedule, and publish social content autonomously. Think of it as the "Stripe for social posting." Your agent handles the content, Typefully handles the delivery.

NOTE

This is a sample gist showcasing how gists.sh renders different file types. Click through the tabs to see Markdown, JSON, YAML, CSV, ICS, and code highlighting in action.

Why AI Agents Need a Publishing Layer

Most AI agents can generate great content. Few can actually publish it. The gap between "here's a draft" and "it's live on 5 platforms" is surprisingly wide: character limits differ, media handling varies, scheduling rules change, and rate limits apply.

Typefully's API abstracts all of that into a single interface.

What You Can Do

Draft and publish to X, LinkedIn, Threads, Bluesky, and Mastodon from one API call
Schedule posts to specific times or Typefully's smart queue slots
Create threads with per-platform formatting
Attach media via presigned upload URLs
Tag and organize drafts for team workflows
Read analytics to close the feedback loop

MCP Integration

Typefully ships an MCP server that works with Claude, Claude Code, and other MCP-compatible agents. Install it once, and your agent can post to social media without any custom API code.

npx typefully-mcp

Or add it to your Claude Code MCP config directly.

Links

nametypefully

descriptionCreate, schedule, and manage social media posts via Typefully

triggers

draft a postschedule a tweetpost to social mediapublish on Xshare on LinkedIncreate a thread

tools

mcp
typefully

Typefully Skill

This is a simplified sample. The real, installable skill is at typefully.com/agent-skills.

When the user asks to create, schedule, or manage social media content, use the Typefully MCP tools.

Workflow

Get the user's social set using typefully_list_social_sets to find available accounts
Check the queue with typefully_get_queue to see existing scheduled posts and open slots
Create the draft with typefully_create_draft, configuring each platform appropriately

Platform-Specific Rules

X/Twitter: 280 character limit per post. Use threads for longer content. Mention with @username.
LinkedIn: Up to 3,000 characters. Tag companies with @[Company Name](urn:li:organization:ID).
Threads: 500 character limit. No link previews.
Bluesky: 300 character limit. Handles auto-detected.
Mastodon: 500 character limit. Use content warnings when appropriate.

Scheduling

Use "next-free-slot" to let Typefully pick the optimal time
Use an ISO 8601 datetime for specific scheduling: "2026-03-15T14:00:00Z"
Use "now" for immediate publishing

Best Practices

Always confirm the content with the user before publishing (publish_at: "now")
Default to saving as draft (omit publish_at) unless the user says "publish" or "post now"
For multi-platform posts, tailor the content per platform rather than copying the same text
Attach the share: true flag when the user wants a preview link

/**
 * publish.ts - Create and schedule a Typefully draft from the command line
 *
 * Usage:
 *   npx tsx publish.ts "Your post content here"
 *   npx tsx publish.ts "Your post content here" --now
 *   npx tsx publish.ts "Your post content here" --schedule 2026-03-15T09:00:00Z
 *
 * Environment:
 *   TYPEFULLY_API_KEY       - your Typefully API key
 *   TYPEFULLY_SOCIAL_SET_ID - your social set ID (find it via the API)
 */

const API_BASE = "https://api.typefully.com/v1";

const apiKey = process.env.TYPEFULLY_API_KEY;
const socialSetId = process.env.TYPEFULLY_SOCIAL_SET_ID;

if (!apiKey) {
  console.error("Error: TYPEFULLY_API_KEY is not set");
  process.exit(1);
}

if (!socialSetId) {
  console.error("Error: TYPEFULLY_SOCIAL_SET_ID is not set");
  process.exit(1);
}

// Parse CLI arguments
const [content, action, scheduleDate] = process.argv.slice(2);

if (!content) {
  console.error('Usage: npx tsx publish.ts "content" [--now|--schedule ISO_DATE]');
  process.exit(1);
}

// Determine publish timing
let publishAt: string | null = null;
if (action === "--now") {
  publishAt = "now";
} else if (action === "--schedule") {
  if (!scheduleDate) {
    console.error("--schedule requires an ISO 8601 date");
    process.exit(1);
  }
  publishAt = scheduleDate;
}

// Create the draft
const response = await fetch(
  `${API_BASE}/social-sets/${socialSetId}/drafts`,
  {
    method: "POST",
    headers: {
      Authorization: `Bearer ${apiKey}`,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({
      platforms: {
        x: {
          enabled: true,
          posts: [{ text: content }],
        },
        linkedin: {
          enabled: true,
          posts: [{ text: content }],
        },
      },
      ...(publishAt && { publish_at: publishAt }),
    }),
  },
);

if (!response.ok) {
  const error = await response.text();
  console.error(`Error (${response.status}):`, error);
  process.exit(1);
}

const draft = (await response.json()) as { id: string };

console.log(`Draft created: #${draft.id}`);

if (publishAt === "now") {
  console.log("Status: Publishing immediately");
} else if (publishAt) {
  console.log(`Status: Scheduled for ${publishAt}`);
} else {
  console.log("Status: Saved as draft");
}

console.log(`View: https://typefully.com/drafts/${draft.id}`);

Fabrizio Rinaldi@linuz90

co-founder typefully.com mailbrew.com (sold) lofi.cafe

Lisbon, Portugalfabrizio.so @linuz90

Platform	Post Type	Impressions	Engagement Rate	Clicks	Reposts	Replies	Best Time	Week
X	Thread	48200	5.8%	1240	312	87	09:00	Feb 17
X	Single Post	12400	3.2%	380	64	23	13:00	Feb 17
X	Quote Post	18900	4.1%	520	145	41	10:00	Feb 17
LinkedIn	Article	8200	6.4%	640	42	38	10:00	Feb 17
LinkedIn	Short Post	5400	4.9%	290	28	15	09:00	Feb 17
Threads	Post	3200	7.1%	180	45	62	15:00	Feb 17
Bluesky	Thread	2800	8.3%	210	88	34	09:00	Feb 17
Bluesky	Post	1600	6.2%	95	41	18	14:00	Feb 17
Mastodon	Post	1200	9.1%	78	32	27	16:00	Feb 17
X	Thread	52100	6.1%	1380	340	95	09:00	Feb 24
X	Single Post	14200	3.5%	410	72	28	13:00	Feb 24
X	Image Post	22400	5.4%	680	198	52	10:00	Feb 24
LinkedIn	Carousel	9800	7.2%	780	56	44	10:00	Feb 24
LinkedIn	Short Post	6100	5.1%	320	31	19	09:00	Feb 24
Threads	Thread	4100	7.8%	240	58	71	15:00	Feb 24
Bluesky	Thread	3400	8.9%	260	102	42	09:00	Feb 24
Mastodon	Thread	1500	9.4%	92	38	31	16:00	Feb 24
X	Thread	45800	5.5%	1180	298	82	09:00	Mar 3
LinkedIn	Article	7600	6.0%	580	38	35	10:00	Mar 3
Threads	Post	3600	7.4%	195	51	58	13:00	Mar 3

/** * publish.ts - Create and schedule a Typefully draft from the command line * * Usage: * npx tsx publish.ts "Your post content here" * npx tsx publish.ts "Your post content here" --now * npx tsx publish.ts "Your post content here" --schedule 2026-03-15T09:00:00Z * * Environment: * TYPEFULLY_API_KEY - your Typefully API key * TYPEFULLY_SOCIAL_SET_ID - your social set ID (find it via the API) */ const API_BASE = "https://api.typefully.com/v1"; const apiKey = process.env.TYPEFULLY_API_KEY; const socialSetId = process.env.TYPEFULLY_SOCIAL_SET_ID; if (!apiKey) { console.error("Error: TYPEFULLY_API_KEY is not set"); process.exit(1); } if (!socialSetId) { console.error("Error: TYPEFULLY_SOCIAL_SET_ID is not set"); process.exit(1); } // Parse CLI arguments const [content, action, scheduleDate] = process.argv.slice(2); if (!content) { console.error('Usage: npx tsx publish.ts "content" [--now|--schedule ISO_DATE]'); process.exit(1); } // Determine publish timing let publishAt: string | null = null; if (action === "--now") { publishAt = "now"; } else if (action === "--schedule") { if (!scheduleDate) { console.error("--schedule requires an ISO 8601 date"); process.exit(1); } publishAt = scheduleDate; } // Create the draft const response = await fetch( `${API_BASE}/social-sets/${socialSetId}/drafts`, { method: "POST", headers: { Authorization: `Bearer ${apiKey}`, "Content-Type": "application/json", }, body: JSON.stringify({ platforms: { x: { enabled: true, posts: [{ text: content }], }, linkedin: { enabled: true, posts: [{ text: content }], }, }, ...(publishAt && { publish_at: publishAt }), }), }, ); if (!response.ok) { const error = await response.text(); console.error(`Error (${response.status}):`, error); process.exit(1); } const draft = (await response.json()) as { id: string }; console.log(`Draft created: #${draft.id}`); if (publishAt === "now") { console.log("Status: Publishing immediately"); } else if (publishAt) { console.log(`Status: Scheduled for ${publishAt}`); } else { console.log("Status: Saved as draft"); } console.log(`View: https://typefully.com/drafts/${draft.id}`);

Typefully Skill

This is a simplified sample. The real, installable skill is at typefully.com/agent-skills.

When the user asks to create, schedule, or manage social media content, use the Typefully MCP tools.

Workflow

Get the user's social set using typefully_list_social_sets to find available accounts
Check the queue with typefully_get_queue to see existing scheduled posts and open slots
Create the draft with typefully_create_draft, configuring each platform appropriately

Platform-Specific Rules

X/Twitter: 280 character limit per post. Use threads for longer content. Mention with @username.
LinkedIn: Up to 3,000 characters. Tag companies with @[Company Name](urn:li:organization:ID).
Threads: 500 character limit. No link previews.
Bluesky: 300 character limit. Handles auto-detected.
Mastodon: 500 character limit. Use content warnings when appropriate.

Scheduling

Use "next-free-slot" to let Typefully pick the optimal time
Use an ISO 8601 datetime for specific scheduling: "2026-03-15T14:00:00Z"
Use "now" for immediate publishing

Best Practices

Always confirm the content with the user before publishing (publish_at: "now")
Default to saving as draft (omit publish_at) unless the user says "publish" or "post now"
For multi-platform posts, tailor the content per platform rather than copying the same text
Attach the share: true flag when the user wants a preview link

Typefully: Social Media Infrastructure for AI Agents

Why AI Agents Need a Publishing Layer

What You Can Do

MCP Integration

Links

Typefully Skill

Workflow

Platform-Specific Rules

Scheduling

Best Practices

Typefully API v2 Launch

MCP Office Hours

Building AI Social Agents Workshop

Typefully Community Meetup - NYC

Why AI Agents Need a Publishing Layer

What You Can Do

MCP Integration

Links

Typefully Skill

Workflow

Platform-Specific Rules

Scheduling

Best Practices