Updated
An AI social media manager that asks before it matters
With Postgale, an AI agent (ChatGPT, Claude, OpenClaw, Hermes Agent or your own) can draft and schedule posts on Instagram, Facebook Pages, Threads, TikTok, Bluesky and Mastodon, read and reply to comments on Instagram, Facebook and Threads, answer direct messages on Instagram, Messenger and WhatsApp, and report on results. You decide where a person must approve: per workspace, account, API key or connected app. Postgale does not yet judge which individual items are sensitive; rules that do that are in early access.
What can the agent do as a social media manager?
Postgale is the secure layer for AI agents on social media. The agent gets 70 tools through the MCP server (or the same operations through the REST API) for the accounts you connected, and nothing else.
| Job | How the agent does it | Where |
|---|---|---|
| Plan and draft | create_post without a time saves a draft; upload_media hosts images and video | Instagram, Facebook Pages, Threads, TikTok, Bluesky and Mastodon |
| Schedule and publish | create_post with scheduledFor, or publish_post; a worker publishes at that time and reports the result | Instagram, Facebook Pages, Threads, TikTok, Bluesky and Mastodon |
| Watch comments | list_comments and list_post_comments; comment.received webhooks | Instagram, Facebook, Threads |
| Reply and moderate | reply_to_comment and moderate_comment (hide, delete) | Instagram, Facebook, Threads |
| Answer DMs | list_conversations, get_conversation_messages, send_message, inside each platform’s messaging window | Instagram, Messenger, WhatsApp |
| Report | get_analytics for post and account insights; tiktok_get_profile and tiktok_list_videos for TikTok stats | Instagram, Facebook, Threads, TikTok |
Comments and direct messages on Instagram and Facebook depend on the permissions Meta grants the connection. If one is missing, the tool answers missing_permission and get_account_health lists what to fix.
Can approval apply only to sensitive items?
Not item by item yet. Today approval is a switch a person sets in the dashboard for the whole workspace, one connected account, one API key or one signed-in app. Where it applies, everything the agent wants to publish, reply or send waits for a person; where it does not, the agent acts directly. Rules that let safe actions through and send only risky ones to the queue (prices, promises, links, quiet hours) are in early access.
You can still put the line where the risk is:
- By account: require approval on the accounts where a mistake costs most (the main Instagram, the WhatsApp number customers write to) and leave it off for, say, a Mastodon announcements account.
- By agent: give an autonomous agent (one that runs on a schedule or from a chat app) a key with approval on, and a person working in Claude or ChatGPT a sign-in without it.
- By habit: ask the agent to save drafts. A post without a time or
publishNowgoes nowhere until someone schedules it. - Always: direct TikTok posts from an agent wait for a person to confirm privacy and interaction settings, whatever the approval settings.
Queued replies and messages are checked before they wait (ownership, the messaging window, validation), so the agent hears about errors at once. A person approves, edits or rejects each item in Dashboard → Approvals; the agent’s key cannot approve its own work.
How do I set it up?
- Create a Postgale workspace and connect the accounts under Dashboard → Accounts.
- Decide where approval applies (workspace, account, key or app) under the dashboard settings.
- Connect the agent: ChatGPT, Claude or Cursor with a sign-in; OpenClaw, Hermes Agent, n8n or your own code with an API key.
- Optional: register a webhook for
comment.received,message.receivedandaction.*events so your own system knows when something needs attention.
Ask your agent“Every morning, check new comments and DMs on our Instagram and Facebook. Draft replies to product questions, leave complaints for me, and draft two posts for this week. Nothing goes out without approval.”
What is available today, and what is in early access?
Available today: publishing and scheduling on Instagram, Facebook Pages, Threads, TikTok, Bluesky and Mastodon; comments on Instagram, Facebook and Threads; direct messages on Instagram, Messenger and WhatsApp; analytics; approval per workspace, account, API key or connected app, with one review queue for posts, replies, messages, moderation and edits to live posts; the audit log; revocable access; the REST API, the MCP server and signed webhooks. X, Telegram and Discord, Reddit and ads are coming soon.
In early access, not generally available: rules checked before every action (so safe actions go through and only risky ones wait), approvals from your phone, client approval links, shadow mode, pause and undo, fine-grained permissions per agent, and a weekly report per client.
Common questions
Yes, within limits you set. Through Postgale it can draft, schedule and publish on Instagram, Facebook Pages, Threads, TikTok, Bluesky and Mastodon, reply to comments on Instagram, Facebook and Threads and answer DMs on Instagram, Messenger and WhatsApp, with a person approving where approval is on.
Not item by item yet: approval applies per workspace, account, API key or connected app. Rules that send only risky actions to the queue are in early access.
No. Postgale handles comments on Instagram, Facebook and Threads. TikTok, Bluesky and Mastodon are publishing only.
Yes. Every write through the API or MCP is recorded in the workspace audit log with the key or app that made it.