Developers

API and MCP server

Let Claude, Cursor, ChatGPT or your own code plan and schedule posts, upload media and hear when posts go out. Every plan includes it, the Free plan too.

1. Make a key

In the app, open Settings, then API and MCP, and press New key. The key starts with cbr_ and shows only once. A key acts as the person who made it, with their role and brands. Delete it any time, and apps using it stop working.

2. Connect an AI agent (MCP)

The MCP server is at https://app.capybaroar.com/api/mcp. Send the key as Authorization: Bearer cbr_... with each request.

Claude Code

claude mcp add --transport http capybaroar https://app.capybaroar.com/api/mcp --header "Authorization: Bearer cbr_..."

Cursor, in ~/.cursor/mcp.json

{
  "mcpServers": {
    "capybaroar": {
      "url": "https://app.capybaroar.com/api/mcp",
      "headers": { "Authorization": "Bearer cbr_..." }
    }
  }
}

Claude, ChatGPT and other apps that take only a link. Add a custom connector with the link that holds your key: https://app.capybaroar.com/api/mcp/cbr_.... Settings shows it ready to copy when you make a key. Treat this link like a password.

Tools

ToolWhat it does
list_brandsThe brands (clients) in the workspace
list_channelsConnected channels with their character limit and the settings each network takes
list_postsPosts in a time range, up to 92 days
get_postOne post, with its status and link on each channel
create_postsUp to 25 posts at once, as drafts unless asked to schedule
update_postChange a post that hasn't gone out
delete_postDelete a post from Capybaroar. With everywhere: true, delete it on the networks too (all its channels, or channel_ids)
list_mediaFiles in the media library, to attach by id
upload_mediaAdd a picture or video from a public link
next_queue_slotThe next free time in the posting queue
get_analyticsReach, engagement, best times to post and top posts
list_webhooks, create_webhook, set_webhook, test_webhook, delete_webhookManage webhooks (owners and admins)

Try asking: "Write a week of LinkedIn posts about our launch, one each weekday at 9 AM, as drafts."

3. REST API

Base address https://app.capybaroar.com/api/v1. Send the key as Authorization: Bearer cbr_... and JSON bodies. Times are ISO 8601 with an offset, like 2026-11-05T09:00:00-05:00.

RequestWhat it does
GET /brandsBrands
GET /channels?brand_id=Channels, with limits and settings
GET /posts?from=&to=&brand_id=Posts in a range (the next 30 days without from and to)
POST /postsMake a post
GET /posts/{id}One post
PATCH /posts/{id}Change the fields you send
DELETE /posts/{id}Delete a post from Capybaroar
DELETE /posts/{id}?everywhere=true&channels=Delete a post on the networks it went out on (all, or the channel ids in channels, comma-separated)
GET /mediaMedia library
POST /mediaAdd a file from a link
GET /queueNext free queue time
GET /analytics?days=30&channel_id=Results for 7, 30 or 90 days
GET /webhooksWebhooks
POST /webhooksAdd a webhook
PATCH /webhooks/{id}Turn it on or off: {"enabled": false}
POST /webhooks/{id}/testSend a ping event now
DELETE /webhooks/{id}Delete a webhook

Make a post

curl https://app.capybaroar.com/api/v1/posts \
  -H "Authorization: Bearer cbr_..." \
  -H "Content-Type: application/json" \
  -d '{
    "channel_ids": ["CHANNEL_ID"],
    "text": "We just shipped the new calendar.",
    "publish_at": "2026-11-05T09:00:00-05:00",
    "mode": "schedule",
    "settings": { "CHANNEL_ID": { "first_comment": "Try it: capybaroar.com" } }
  }'
FieldWhat it is
channel_idsChannels from one brand. Required.
textThe text every channel shares
modedraft (the default), schedule, now or queue
publish_atRequired for draft and schedule
media_idsFiles from the media library, in order
threadA thread's posts after the first, each { "text", "media_ids" }, up to 24. They go out as replies on X, Bluesky, Threads and Mastodon. Other channels post the first only.
channel_textOwn text for some channels, by channel id
settingsNetwork settings by channel id, like a first comment, a YouTube title or a Pinterest board

Posts get the same checks as in the app: character limits, required settings, media, plan limits, brands and approvals. In a brand that needs approval, an editor's post waits in review.

Delete a post

A plain delete removes the post from Capybaroar only: what already went out stays on the networks. With everywhere, each channel is deleted on its network and the answer lists what happened on each one. The post stays in Capybaroar with the status deleted once every channel is, so analytics keep its results; list_posts leaves it out. Delete it again without everywhere to remove the record.

curl -X DELETE "https://app.capybaroar.com/api/v1/posts/POST_ID?everywhere=true" \
  -H "Authorization: Bearer cbr_..."

{
  "deleted": false,
  "channels": [
    { "channel_id": "...", "name": "Acme", "network": "linkedin-page", "result": "deleted", "error": null, "note": null, "url": "https://www.linkedin.com/feed/update/..." },
    { "channel_id": "...", "name": "acme", "network": "instagram", "result": "manual", "error": "Delete it in the Instagram app", "note": null, "url": "https://www.instagram.com/p/..." }
  ]
}

result is deleted, failed (with the error; try again), manual (the network can't delete from here: open url and delete it there; Instagram and TikTok always) or skipped (nothing went out on that channel). deleted is true once every channel is deleted.

Upload a file by link

Send a public link and Capybaroar downloads the file into the media library. It takes JPG, PNG, GIF, WebP, MP4 and MOV up to 250 MB, judged by the file itself. Links to private networks are refused.

curl https://app.capybaroar.com/api/v1/media \
  -H "Authorization: Bearer cbr_..." \
  -H "Content-Type: application/json" \
  -d '{ "url": "https://example.com/launch.png", "alt": "The new calendar" }'

The answer has the file's id. Put it in media_ids when you make a post.

Webhooks

Add a webhook in Settings or with POST /webhooks, with an https URL and the events you want. Capybaroar sends a POST to it when they happen.

EventWhen
post.publishedA post went out on a channel, with its link
post.failedA post failed on a channel after its retries, with the reason
post.in_reviewA post waits for approval
post.approvedA post was approved
post.rejectedA post was sent back, with the comment
channel.reconnectA channel needs reconnecting
POST https://your-site.com/hooks/capybaroar
Capybaroar-Event: post.published
Capybaroar-Delivery: evt_6f1c...
Capybaroar-Signature: t=1793700000,v1=5257a8...

{
  "id": "evt_6f1c...",
  "type": "post.published",
  "created_at": "2026-11-05T14:00:03.120Z",
  "workspace_id": "...",
  "data": { "post_id": "...", "brand_id": "...", "channel_id": "...", "channel_name": "Perceptric", "network": "linkedin", "url": "https://www.linkedin.com/feed/update/..." }
}

Check the signature. You get the signing secret once, when you add the webhook. Check each request before you trust it:

import { createHmac, timingSafeEqual } from "node:crypto";

function verify(secret, header, rawBody) {
  const { t, v1 } = Object.fromEntries(header.split(",").map((p) => p.split("=")));
  const want = createHmac("sha256", secret).update(`${t}.${rawBody}`).digest("hex");
  const fresh = Math.abs(Date.now() / 1000 - Number(t)) < 300;
  return fresh && timingSafeEqual(Buffer.from(want), Buffer.from(v1));
}

Answer with a 2xx within 10 seconds. Anything else is tried again after 1, 5, 15 and 60 minutes, then 3, 6 and 12 hours. A webhook whose deliveries fail 5 times in a row turns itself off, and Settings shows why. Every try of an event has the same id, so you can skip ones you already handled.

Errors and limits

Errors come back as {"error": {"code": "...", "message": "..."}} with a status: 400 for a request to fix, 401 for a missing or deleted key, 404 when something isn't there, 409 when a post already went out, 429 when you go over 120 requests a minute (wait for the seconds in Retry-After).