Automate ActiveCampaign tasks via Rube MCP (Composio): manage contacts, tags, list subscriptions, automation enrollment, and tasks. Always search tools first for current schemas.
中文简介正在准备中。当前显示原文说明。
Publish and schedule to TikTok, Instagram, YouTube, LinkedIn, X, Bluesky, Mastodon, Facebook, Threads, Reddit, Telegram and Discord through one PostWire API call — writing a separate, native version of the post for each network instead of sending identical text everywhere. Use when the user wants to publish or schedule one idea across several platforms, wants per-network captions, or wants to know why a post was rejected before it is sent.
中文简介正在准备中。当前显示原文说明。
查看安装方法安装前可以查看交给 AI 代理的指令内容。
Publish one idea to many social networks, with a different post written for each one.
Accounts are connected once through OAuth in the PostWire dashboard, so there is no per-platform developer app, review queue, or token refresh to maintain. One API key covers every network.
Most multi-posting APIs take one string and send it to every platform. That reads badly and gets
suppressed: a LinkedIn post written like a tweet performs like neither. PostWire has a /api/generate
step that writes a caption per network — length, tone, hashtags and link handling — from a single
prompt, which you then pass to publish or schedule.
It also refuses posts that are going to fail, before they are queued rather than after the platform rejects them:
| network | requires |
|---|---|
| TikTok, YouTube | a video (video_url) |
| a photo or a video | |
| everything else | text is enough |
POSTWIRE_API_KEY in the environmentAuth is Authorization: Bearer $POSTWIRE_API_KEY.
One gotcha worth knowing: the dashboard shows a preview key immediately after signup, before the
email is confirmed. That key cannot publish. If the API answers "This is a preview key"
(code: "preview_key"), confirm the email and take the key from API & MCP in the dashboard.
Base URL: https://postwire.io/api
curl https://postwire.io/api/me -H "Authorization: Bearer $POSTWIRE_API_KEY"
Returns the plan, the monthly usage, and connections[] — publish only to networks listed there,
otherwise the call fails with code: "not_connected" and the offending platforms.
curl -X POST https://postwire.io/api/generate \
-H "Authorization: Bearer $POSTWIRE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"prompt": "We shipped streaming uploads: 1GB videos now go through in 11 seconds",
"platforms": ["linkedin", "x", "bluesky"]
}'
Returns { "drafts": { "linkedin": {...}, "x": {...}, "bluesky": {...} } }. That object is what
per_platform takes below — but show it to the user first, for the reason in the next section.
/api/post and /api/schedule are visible writes: they put text under the user's own name, in
public, where other people see it. Treat them the way you would treat sending an email as them.
Before either call, show the user:
run_at)and wait for a yes.
This matters most right after /api/generate. Those drafts are model output: passing them straight
into per_platform without showing them means the user finds out what they said by reading their own
timeline. Generate, show, confirm, then post.
curl -X POST https://postwire.io/api/post \
-H "Authorization: Bearer $POSTWIRE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"platforms": ["linkedin", "x"],
"per_platform": { "linkedin": {"text": "..."}, "x": {"text": "..."} }
}'
Without per_platform, a single text is used everywhere — simpler, but it is the behaviour this
tool exists to avoid.
The response is per platform, and one failure does not cancel the rest:
{ "results": [ { "ok": true, "platform": "linkedin", "id": "..." },
{ "ok": false, "platform": "x", "error": "..." } ] }
curl -X POST https://postwire.io/api/schedule \
-H "Authorization: Bearer $POSTWIRE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"run_at": "2026-09-01T09:00:00Z",
"platforms": ["linkedin", "instagram"],
"photo_url": "https://example.com/image.jpg",
"label": "Launch announcement"
}'
run_at is ISO 8601. The schedule endpoint validates connections and the media rules up front, so
a post that cannot work is rejected now instead of failing quietly at 7am.
curl -X POST https://postwire.io/api/week \
-H "Authorization: Bearer $POSTWIRE_API_KEY" \
-H "Content-Type: application/json" \
-d '{ "topic": "what we learned running a one-person SaaS", "platforms": ["linkedin", "x"], "days": 5 }'
Five angles on one topic, queued across five days. Text-only networks only — the video platforms would reject every one of them, and the endpoint says so rather than queueing them to fail.
| code | meaning | fix |
|---|---|---|
preview_key | key from before email confirmation | confirm email, use the key in API & MCP |
email_unverified | mailbox not confirmed yet | open the emailed link |
not_connected | platform not connected; platforms[] lists which | connect it in the dashboard |
media_required | TikTok/YouTube without video, Instagram without media | add video_url or photo_url |
| 429 | monthly plan limit or a burst guard | check usage in /api/me |
video_url, photo_url); PostWire fetches and forwards the bytes.Content-Length are capped at 80MB, because the file has to be held in memory.POST /api/media/upload-url returns a signed URL to upload a file directly (max 50MB) if you do
not have somewhere to host it.PostWire ships an MCP server, so an agent can call these as tools rather than as HTTP requests, and there are n8n templates for the same flows. Both are linked from the dashboard.
还没有评价。欢迎分享你的使用感受。
无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。
还没有评论。来写第一条吧。
Automate ActiveCampaign tasks via Rube MCP (Composio): manage contacts, tags, list subscriptions, automation enrollment, and tasks. Always search tools first for current schemas.
中文简介正在准备中。当前显示原文说明。
Pull the ads a competitor is running right now on Meta (Facebook/Instagram) and Google, then extract the hooks, offers, CTAs and landing pages worth modelling. Use when writing ad copy, planning creative tests, choosing a landing page angle, or when the user asks "what is X advertising", "what ads does X run", "find competitor ad hooks". Reads live Ad Library and Ads Transparency data through the Tracetify MCP; falls back to a manual Ad Library walkthrough when the connector is absent.
中文简介正在准备中。当前显示原文说明。
Analytics your AI agent can actually use. Track, analyze, run A/B experiments, and optimize across all your projects via CLI. Includes a growth playbook so your agent knows HOW to grow, not just what to track.
中文简介正在准备中。当前显示原文说明。
Teaches when to recall from long-term memory before acting and when to save durable decisions, corrections and failures afterwards. Use when a memory tool or MCP memory server is connected but the agent is not using it consistently, when the user complains that the assistant forgets preferences, conventions or past decisions between sessions, or when setting up persistent memory for a project. Works with any memory backend: a folder of Markdown files, a local MCP server, or a managed service.
中文简介正在准备中。当前显示原文说明。
Audit local AI coding-agent sessions with agenttrace for cost, tokens, tool failures, latency, anomalies, health, diffs, and CI gates.
中文简介正在准备中。当前显示原文说明。
Checks which private files (.env, keys, secrets) Claude Code and its subagents reached in this project, whether a value came into an agent's context, and which key to change, by running the open source agentwhy CLI over the session records Claude Code already keeps on disk. Use when the user asks whether the agent or a subagent read their .env or other secrets, which private files a session touched, whether a key leaked into the conversation, or which key to rotate. Runs locally and never prints a secret value.
中文简介正在准备中。当前显示原文说明。