PostWire API reference

Publish one idea as a native post on each network, schedule it, and get told when it goes out. JSON over HTTPS at https://postwire.io. Every endpoint has a copyable example in curl, Node and Python.

View as MarkdownOpenAPIBase URL https://postwire.io

Quick start

1. Get a key. Sign up at postwire.io and open API & MCP. Keys start with pw_live_. The Free plan includes the API.
2. Connect a network. Telegram, Bluesky, Mastodon and Discord with POST /api/connect; TikTok, YouTube, Instagram, Facebook Pages and LinkedIn with OAuth (POST /api/oauth/{platform}/url). Or in the dashboard.
3. Publish. One request, every network you name:

curl -X POST "https://postwire.io/api/post" \
  -H "Authorization: Bearer $POSTWIRE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "platforms": [
    "bluesky",
    "telegram"
  ],
  "text": "Hello from the PostWire API"
}'

const res = await fetch("https://postwire.io/api/post", {
  method: "POST",
  headers: { Authorization: `Bearer ${process.env.POSTWIRE_API_KEY}`, "Content-Type": "application/json" },
  body: JSON.stringify({
    "platforms": [
      "bluesky",
      "telegram"
    ],
    "text": "Hello from the PostWire API"
  }),
});
console.log(res.status, await res.json());

import os, requests

r = requests.post(
    "https://postwire.io/api/post",
    headers={"Authorization": f"Bearer {os.environ['POSTWIRE_API_KEY']}"},
    json={
        "platforms": ["bluesky", "telegram"],
        "text": "Hello from the PostWire API",
    },
)
print(r.status_code, r.json())

Authentication

Send the key in Authorization: Bearer pw_live_… (or X-API-Key: pw_live_…). A missing or wrong key answers 401. Publishing needs a verified email and a full-access key, otherwise 403. If PostWire cannot check a key because of its own outage it answers 503 auth_unavailable, never 401: retry, do not throw the key away.

Keys can be created with an expiry and revoked at once: see Account, keys and billing. GET /api/platforms, GET /api/plans and GET /api/status need no key.

One result per network

Publishing answers 200 with posted (networks that went out) and results[], one entry per network, each with ok and, when it did not go out, error and a machine-readable code. One network failing never fails the request or the others. A request-level problem (bad JSON, nothing connected, a network connected in two brands without brand_id) answers 4xx before anything is sent.

A post counts once per network that published: one idea to three networks is three of the month's posts. Networks refused before sending (not connected, held by the plan, media missing) spend nothing.

Pages

Plans and limits

From GET /api/plans, the same table the billing code reads. A post counts once per network. Every paid plan can be billed monthly or yearly at 20% off (interval: "year" on POST /api/billing/checkout, ?interval=year on GET /api/billing/upgrade-link).

Every plan includes the REST API, the MCP server, webhooks and TikTok. See pricing.

Fair use: every limit

"Unlimited posts" on Pro, Agency and Scale has a published ceiling that no honest customer meets, and these are the only limits. They are per workspace, enforced on the server, and each one answers with its own code, the number and when it resets (see errors and limits). GET /api/usage and GET /api/me show today's posts and the month's media against them.

A sudden spike — more posts in one hour than 10 times the workspace's usual hour, and never less than a quarter of its day — pauses publishing until the hour ends (throttled); scheduled posts keep their place. Media is kept 30 days. X posts use X credits on top of these.

Machine-readable

• OpenAPI 3.0: /openapi.json.
• Every page of this reference as Markdown: add .md (for example /docs/publishing.md), or use the Markdown link at the top of each page.
• MCP server for Claude, ChatGPT and other agents: /mcp/. For AI crawlers: /llms.txt.