GetMulticast API

Publish, schedule, and manage replies across every platform GetMulticast connects to — programmatically. Built for Zapier, n8n, Make, and your own scripts.

This is the same engine the GetMulticast dashboard itself runs on, not a separate or lighter reimplementation — every endpoint here calls the identical internal logic a click in the dashboard would trigger, so behavior never drifts between the two. A UI redesign never changes what these endpoints do; only a real, documented change to this reference does.

Official SDKs: none published yet — JavaScript, Python and PHP clients are planned, not built. Until then, any HTTP client works fine: every example on this page is plain curl, and translates directly to fetch, requests, axios, Guzzle, or whatever you already use.

Quickstart

Zero to a real published post, in order:

1. Get an API key — Dashboard → Settings → Developer. This one step is UI-only; everything after it is a plain HTTP call.
2. Find a real account id — every id used below ("1784..." etc.) comes from this call, not made up:

curl https://getmulticast.com/api/v1/accounts \
  -H "Authorization: Bearer gmc_live_..."

3. Publish — using a real mediaUrl you already have hosted somewhere public, and a real account id from step 2:

curl -X POST https://getmulticast.com/api/v1/posts \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "mediaUrl": "https://example.com/clip.mp4",
    "mediaType": "video",
    "caption": "Hello from the API",
    "selections": { "instagram": ["1784..."] }
  }'

That returns 202 immediately with a runId — see Rate limits & timeouts for why it's 202 and not 201, and GET /posts/{id} below for how to check the real result once publishing finishes.

Authentication

Every request needs an API key in the Authorization header, as a Bearer token. Create one from Dashboard → Settings → Developer — the raw key is shown exactly once at creation and can't be retrieved again (we only ever store a one-way hash of it, the same way Stripe and GitHub handle API keys).

# every request
curl https://getmulticast.com/api/v1/accounts \
  -H "Authorization: Bearer gmc_live_XXXXXXXXXXXXXXXXXXXXXXXX"

Each key carries its own scopes, checked directly against this real list — nothing here is aspirational:

Scope	Grants
accounts:read	Read connected social accounts
posts:read	Read posts, publish results and scheduled posts
posts:write	Publish, schedule, reschedule, delete and retry posts
analytics:read	Read real per-post metrics (YouTube, X, Pinterest, Bluesky only — see GET /posts/{id}/analytics)
comments:read	Read recent comments and the pending-reply review queue
comments:write	Draft, edit, approve, reject and send comment replies
credits:read	Read the account's AI credit balance
webhooks:read	List registered webhooks and their delivery history
webhooks:write	Register and delete webhooks
rss:read	List RSS auto-post feeds
rss:write	Register, update and delete RSS auto-post feeds
recycling:read	List content recycling rules
recycling:write	Create, update and delete content recycling rules

A request to an endpoint outside a key's granted scopes returns 403 insufficient_scope. Create narrowly-scoped keys for anything you hand to a third-party tool.

Rate limits & timeouts

Limit	Value
Requests per minute, per key	60
Requests per day, per key	5,000
Request timeout	30s
Max request body	10MB

Every response carries rate-limit headers so you can self-throttle before hitting the ceiling:

X-RateLimit-Limit: 60
X-RateLimit-Remaining: 57
X-RateLimit-Reset: 1757980800   # unix epoch seconds

Exceeding a limit returns 429 with a Retry-After header (seconds to wait). POST /posts and POST /posts/scheduled queue the real publish work in the background and return immediately (202) — the actual per-platform publishing (which can genuinely take minutes, e.g. Instagram/TikTok video processing) happens after the response, same as the dashboard's own "Post Now" button. Poll GET /posts/{id} for the final result, or use a webhook (see below) instead of polling.

Errors

Every error — auth, validation, rate limit, or an unexpected failure — returns the same JSON envelope, so you can always branch on error.code rather than parsing text:

{
  "error": {
    "code": "insufficient_scope",
    "message": "This API key doesn't have the 'posts:write' scope required for this endpoint.",
    "requestId": "a1b2c3d4e5f6a1b2"
  }
}

HTTP status	error.code	Meaning
400	invalid_request	Missing/malformed field — see message for which one.
401	missing_api_key / invalid_api_key / revoked_api_key	No, wrong, or revoked Bearer token.
403	insufficient_scope / forbidden	Key lacks the scope, or the resource belongs to someone else.
404	not_found	No such resource, or wrong method for a real path.
405	method_not_allowed	Path exists, wrong HTTP method.
409	failed_precondition	Valid request, wrong state (e.g. approving an already-sent reply).
429	rate_limit_exceeded / resource_exhausted	Rate limit hit, or (for posts) out of AI credits.
500	internal_error	Something broke on our end — logged server-side by requestId; include it if you contact support.

Two real mistakes worth calling out

An empty selections object on POST /posts is rejected up front, before anything is queued:

{ "error": { "code": "invalid_request", "message": "selections is required", "requestId": "..." } }

A misspelled or unsupported platform key inside selections (e.g. "instagrm", or a real platform you just haven't connected yet) behaves differently, and it's easy to get caught out by this: it is not rejected up front. The request still returns 202 like any other, but that platform's entry in the eventual publishResults (see GET /posts/{id}) comes back as a real, specific per-platform failure instead of a request-level error:

"publishResults": {
  "instagrm": [{ "ok": false, "error": "No connected instagrm account found." }]
}

If a platform key never shows up publishing successfully, check it against GET /accounts first — that's the exact, real set of platform keys and account ids you have to work with.

Pagination

List endpoints accept limit (1–100, default 25) and cursor, and return a nextCursor — pass it back as cursor to fetch the next page; null means you're on the last page.

curl "https://getmulticast.com/api/v1/posts?limit=50&cursor=abc123" \
  -H "Authorization: Bearer gmc_live_..."

Idempotency

Send an Idempotency-Key header on POST /posts, POST /posts/scheduled, or POST /comments/reply to make a retried request safe — if you send the same key twice (e.g. your client timed out and retried), the second call returns the exact same result as the first instead of publishing/replying twice. Keys are scoped per API key and remembered for 24 hours. Use a UUID you generate per logical action, same convention as Stripe.

curl -X POST https://getmulticast.com/api/v1/posts \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Idempotency-Key: 8f14e45f-ceea-4f9f-9b0f-3c1e2e6a1a11" \
  -H "Content-Type: application/json" \
  -d '{ ... }'

Versioning

This is v1 — the path itself (/api/v1/...) is the version. Once you're integrated, this contract won't change under you: no field is removed or repurposed, no endpoint's meaning changes. Real new capability ships as new, additive fields or endpoints within v1 where possible; a genuine breaking change ships as /api/v2/... alongside v1, not in place of it.

Accounts

Every connected social account across every platform.

curl https://getmulticast.com/api/v1/accounts \
  -H "Authorization: Bearer gmc_live_..."

{
  "accounts": {
    "instagram": [{ "id": "1784...", "label": "Main Instagram" }],
    "youtube": [{ "id": "UCxx...", "label": "My Channel" }],
    "x": [{ "id": "14582...", "label": "@handle" }]
    // ...one array per connected platform: facebook, threads, tiktok,
    // linkedin, bluesky, pinterest, googlebusiness
  },
  "requestId": "..."
}

Posts

Publish now to one or more connected accounts across one or more platforms. Returns 202 immediately with a post id — publishing itself happens in the background (see Rate limits above). Upload your media somewhere public first (or use mediaUrl from your own storage) — this endpoint never accepts raw file bytes.

Field	Type	Required	Notes
mediaUrl	string	yes	A public URL to the video or image file.
mediaType	string	yes	"video" or "image". Text-only posts aren't supported yet.
selections	object	yes	{ "instagram": ["accountId1"], "x": ["accountId2"] } — platform → array of account ids from GET /accounts.
caption	string	no	Shared caption for every selected account.
captionOverrides	object	no	{ "x": "Shorter caption for X" } — per-platform override.
hashtags	string[]	no	Appended to the caption.
firstComment	string	no	Auto-posted as the first comment (Facebook & Instagram only today).

curl -X POST https://getmulticast.com/api/v1/posts \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "mediaUrl": "https://example.com/clip.mp4",
    "mediaType": "video",
    "caption": "New drop 🚀",
    "hashtags": ["launch","newproduct"],
    "selections": { "instagram": ["1784..."], "youtube": ["UCxx..."] }
  }'

{ "runId": "8fJ2k...", "queued": true, "requestId": "..." }

Lists your posts, newest first. Paginated (see Pagination above).

{
  "posts": [
    { "id": "8fJ2k...", "status": "published", "caption": "New drop 🚀",
      "publishResults": { "instagram": [{ "accountId": "1784...", "ok": true, "postId": "..." }] },
      "createdAt": "2026-09-12T10:03:00Z" }
  ],
  "nextCursor": null,
  "requestId": "..."
}

One post's full status and per-account publish results. status is processing, published, or publish_failed.

Real, live metrics for one post on one platform. Only youtube, x, pinterest, and bluesky are wired — every other platform returns a real 400 not_supported naming the exact missing permission (e.g. Instagram needs a permission this app doesn't have granted yet), never a fake zero.

curl "https://getmulticast.com/api/v1/posts/8fJ2k.../analytics?platform=x" \
  -H "Authorization: Bearer gmc_live_..."

{
  "platform": "x", "postId": "18472...",
  "metrics": { "likes": 42, "reposts": 5, "replies": 3, "quotes": 0, "impressions": 1204 },
  "requestId": "..."
}

Totals (views, interactions, posts, engagement rate) and your top 3 posts by views, across one channel or every connected channel. Both params optional — omit accountId to summarize everything; days defaults to 30 (max 90).

curl "https://getmulticast.com/api/v1/analytics/summary?days=7" \
  -H "Authorization: Bearer gmc_live_..."

One channel's own totals, plus which content formats and hashtags actually perform (medians, not averages, so one viral post doesn't skew it).

This channel's best times/day/format/posting-frequency to post, computed from its own real post history. Returns "basedOn": "general" with generic guidance if there isn't 5+ posts of real history yet — never presented as personalized when it isn't.

Your best-performing posts, ranked by views (default), likes, or comments. All params optional.

Every competitor you're tracking (Analytics → Competitor tracking), with their latest followers, posting frequency, and engagement per post.

Retries only the accounts that failed on a given platform for this post.

curl -X POST https://getmulticast.com/api/v1/posts/8fJ2k.../retry \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "platform": "instagram" }'

Same fields as POST /posts, plus:

Field	Type	Required	Notes
date	string	yes	"2026-09-20"
time	string	no	"09:00", defaults to 09:00.
timezone	string	no	IANA name, e.g. "America/New_York". Defaults to UTC.

Returns 201 with the scheduled post's id and resolved UTC time.

Lists posts still waiting to publish, soonest first. Paginated.

Reschedules a still-pending post. Body: { "scheduledAtUtc": "2026-09-21T14:00:00Z" }, or { "date": "...", "time": "..." } in the post's original timezone.

Cancels a still-pending scheduled post.

Comments & Replies

Recent comments (last 7 days) across your monitored accounts on Instagram, Facebook, Threads, YouTube, LinkedIn and Bluesky, including whether each already has a reply. X/Twitter works at the mentions level instead — see /replies/pending, which covers every platform including X.

Replies to one comment surfaced by GET /comments. Send directText to send your own exact text immediately (no AI, no review step), or omit it to have GetMulticast's AI draft a reply into your Review queue (/replies/pending) instead of sending it directly.

Field	Required	Notes
platform, accountId, commentId	yes	From GET /comments.
text	yes	The original comment's text (used for the AI draft).
directText	no	Send exactly this, no AI involved.
instruction	no	One-off AI instruction, e.g. "mention our discount code".
postUrn	LinkedIn only	Required for LinkedIn.
rootUri, rootCid, commentCid	Bluesky only	Required for Bluesky (AT Protocol reply refs).

Marks a comment as handled without replying (e.g. spam). Body: { "platform", "accountId", "commentId" }.

Your Review Inbox — AI-drafted replies (from review-mode automations, across every platform including X) waiting for approval.

Sends the pending reply — the AI draft as-is, or pass { "editedText": "..." } to send different text instead.

Discards a pending reply without sending it.

Asks the AI to revise the current draft. Body: { "context": "make it shorter and friendlier" } (required).

Overwrites the draft text without sending it. Body: { "text": "..." }.

Credits

Your remaining AI credit balance — spent by POST /videos below (never by the post/comment/webhook/RSS/recycling endpoints above, which only move media you already have).

{ "credits": 847, "requestId": "..." }

AI Video Generation

Generates a real AI video — the same engine the dashboard's own "Generate Video" button calls, not a separate or lighter version. A short text-to-video clip is generated per scene (Agnes), with an optional AI-narrated voiceover track. Generation happens in the background and can take minutes: POST /videos returns 202 immediately with a videoId; poll GET /videos/{id} until status is "ready" (then videoUrl is set — pass it straight into POST /posts or POST /posts/scheduled to publish it) or "failed" (statusMessage explains why; any charged credits are automatically refunded). Honest current limitation: there is no compositing step — with multiple scenes, the first scene that produced a clip becomes videoUrl, every scene's own clip is kept on scenes[i].mediaUrl, and a voiceover track (if requested) is a separate file, not muxed into the video.

Starts generating a video. Charges AI credits immediately (see GET /credits) unless saveAsDraft is true — an insufficient balance returns 429 resource_exhausted before anything is charged or generation starts.

Field	Type	Required	Notes
prompt	string	no	What the video should be about — becomes a single implicit scene when scenes is omitted. Needed unless every scene in scenes supplies its own aiPrompt.
scenes	array	no	Advanced: multiple scenes instead of one prompt. Each: { "mediaSource": "ai"|"upload", "aiPrompt", "mediaUrl", "voiceoverScript" } — "upload" scenes need mediaUrl, "ai" scenes need aiPrompt or a top-level prompt.
aspectRatio	string	no	"9:16" (default), "1:1", or "16:9".
aiVoiceover	boolean	no	Adds an AI-narrated voiceover track (edge-tts), synthesized from each scene's voiceoverScript or the top-level prompt.
voiceId	string	no	Voiceover voice id. Defaults to the account's default voice.
saveAsDraft	boolean	no	Saves the video's settings without generating or charging credits — generate later from the dashboard.

curl -X POST https://getmulticast.com/api/v1/videos \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A cinematic drone shot flying over a mountain lake at sunrise",
    "aspectRatio": "9:16",
    "aiVoiceover": true
  }'

{ "videoId": "vX9k2...", "creditsRemaining": 823, "requestId": "..." }

Lists your videos, newest first. Paginated (see Pagination above).

One video's generation status. status is draft, processing, ready, or failed; progressPct and statusMessage update live while processing.

{
  "id": "vX9k2...", "status": "ready", "progressPct": 100,
  "videoUrl": "https://storage.googleapis.com/.../vX9k2....mp4",
  "voiceoverUrl": "https://storage.googleapis.com/.../vX9k2....mp3",
  "requestId": "..."
}

Permanently deletes a video.

Webhooks

Register a URL to get a real HTTP POST the moment something happens — no polling. Delivery is durable: every event is queued and a background worker attempts delivery, retrying on a real backoff schedule (1m → 5m → 30m → 2h → 6h → 24h, 7 attempts total) if your endpoint doesn't return a 2xx — after which the delivery is marked failed and stops retrying. Every delivery is signed so you can verify it actually came from GetMulticast, and carries the same X-GetMulticast-Delivery id across every retry attempt so you can safely dedupe on your end.

Event	Fires when
post.published	A post finished publishing to at least one account.
post.failed	A post failed on every selected account.
comment.received	A new comment/mention was fetched on a monitored account.
reply.sent	A reply — auto-sent or approved from the Review queue — actually went out.

// POST to your url, Content-Type: application/json
{
  "id": "a5e0...",
  "event": "post.published",
  "createdAt": "2026-09-12T10:04:12Z",
  "data": { "runId": "8fJ2k...", "results": { "...": "..." }, "okCount": 2, "total": 2 }
}

Headers on every delivery:

X-GetMulticast-Event: post.published
X-GetMulticast-Signature: sha256=<hmac hex digest of the raw body, using your webhook secret>
X-GetMulticast-Delivery: a5e0f21c-...   # unique per delivery attempt

Verify it (Node example — same HMAC-SHA256-over-raw-body pattern Stripe/GitHub webhooks use):

const crypto = require("crypto");
function verify(rawBody, signatureHeader, secret) {
  const expected = "sha256=" + crypto.createHmac("sha256", secret).update(rawBody).digest("hex");
  return crypto.timingSafeEqual(Buffer.from(signatureHeader), Buffer.from(expected));
}

Registers a webhook. Body: { "url": "https://...", "events": ["post.published","reply.sent"] }. The response's secret field is shown exactly once — store it to verify future deliveries.

Lists your registered webhooks (never returns the secret again).

Removes a webhook.

Real delivery history for one webhook, newest first — every attempt's status and error, useful for debugging a receiver that isn't seeing events.

{
  "deliveries": [
    { "id": "d9c1...", "event": "post.published", "status": "delivered", "attempts": 1, "lastError": null, "createdAt": "..." },
    { "id": "a2e0...", "event": "reply.sent", "status": "pending", "attempts": 2, "lastError": "http_503", "nextAttemptAt": "..." }
  ],
  "requestId": "..."
}

RSS Auto-Post

Point GetMulticast at an RSS or Atom feed and it checks for new items on a schedule (15 minutes minimum), posting each new one to the accounts you choose — the same automation Zapier/IFTTT "RSS → social" recipes provide, built in.

Field	Required	Notes
url	yes	The feed's URL (RSS 2.0 or Atom).
selections	yes	Same shape as POST /posts — { "instagram": ["accountId"] }. Only image-capable platforms are accepted.
captionTemplate	no	Defaults to "{title}\n\n{link}" — {title} and {link} are the only placeholders substituted.
checkIntervalMinutes	no	Defaults to 60, floor of 15.

curl -X POST https://getmulticast.com/api/v1/rss-feeds \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/feed/",
    "selections": { "instagram": ["1784..."], "facebook": ["1029..."] },
    "captionTemplate": "New post: {title}\n\nRead more: {link}",
    "checkIntervalMinutes": 30
  }'

Lists your feeds, including lastCheckedAt, lastError (if the last check failed) and postedCount.

Updates any of active, selections, captionTemplate, checkIntervalMinutes. Set active: false to pause a feed without deleting it.

Queues an immediate check instead of waiting for the next scheduled interval — picked up within about a minute.

Content Recycling

Pick one of your own already-published posts and set up a rule to automatically re-publish it again every N days — optionally with the caption AI-rewritten each time so it doesn't read as a verbatim repeat. Checked every 6 hours.

Field	Required	Notes
sourcePostId	yes	A post id from GET /posts that already published successfully.
selections	no	Defaults to the accounts the source post actually succeeded on.
intervalDays	no	Defaults to 30, floor of 1.
rewriteCaption	no	If true, AI rewrites the caption on each recycle instead of reusing it verbatim.

curl -X POST https://getmulticast.com/api/v1/recycling-rules \
  -H "Authorization: Bearer gmc_live_..." \
  -H "Content-Type: application/json" \
  -d '{ "sourcePostId": "8fJ2k...", "intervalDays": 14, "rewriteCaption": true }'

Lists your rules, including lastRecycledAt, nextRecycleAt, and recycleCount.

Updates any of active, selections, intervalDays, rewriteCaption. Set active: false to pause without deleting.

Queues an immediate recycle instead of waiting for the next scheduled interval.

OpenAPI spec & Postman collection

Every endpoint on this page, machine-readable — generate a client in any language, or import straight into Postman/Insomnia instead of building requests by hand.

Download OpenAPI 3.0 spec (.yaml)
Download Postman collection (.json)

The collection ships with base_url and api_key as collection variables — set api_key once (Postman → collection → Variables) and every request in it authenticates automatically.

MCP server

A Model Context Protocol server that exposes this entire API as tools an MCP-compatible AI client (ChatGPT, Claude, etc.) can call directly — "post this to Instagram and Facebook," "what's in my review queue," "reply to that comment," said in plain language. Two ways to connect, use whichever fits your client — GetMulticast never sees your MCP client's traffic either way, it's just another caller of the same public API documented on this page.

Remote server (no install) — for ChatGPT, Claude Code, and claude.ai

Server URL: https://getmulticast.com/mcp — every request needs your API key as a normal bearer header: Authorization: Bearer gmc_live_...

ChatGPT (Settings → Connectors → Advanced → Developer mode, then Add MCP server):

• Transport: Streamable HTTP
• URL: https://getmulticast.com/mcp
• Headers: name Authorization, value Bearer gmc_live_...

Claude Code, one command, no download:

claude mcp add getmulticast -t http \
  --header "Authorization: Bearer gmc_live_..." \
  -- https://getmulticast.com/mcp

claude.ai (web or desktop app) — no API key needed, real sign-in instead:

1. Settings → Connectors → Add custom connector
2. URL: https://getmulticast.com/mcp
3. Authentication: Sign in now
4. Click Add — Claude opens a GetMulticast login/approve screen, click Approve, done

Local server (download & run) — Claude Desktop's other option

Download the server (.zip) — current version: v1.3.0 · Released Sept 2026 — unzip, then:

cd getmulticast-mcp-server
npm install
GETMULTICAST_API_KEY=gmc_live_... node index.js   # test it starts cleanly

Then point your MCP client at it. For Claude Desktop, add to its config file (Settings → Developer → Edit Config):

{
  "mcpServers": {
    "getmulticast": {
      "command": "node",
      "args": ["/full/path/to/getmulticast-mcp-server/index.js"],
      "env": { "GETMULTICAST_API_KEY": "gmc_live_..." }
    }
  }
}

Claude Code can use this local server too, if you'd rather run it yourself than use the remote URL above:

claude mcp add getmulticast \
  --env GETMULTICAST_API_KEY=gmc_live_... \
  -- node /full/path/to/getmulticast-mcp-server/index.js

Same rule applies — replace /full/path/to/ with your own real unzip location, same as above.

37 tools total, one per API endpoint on this page, verified against the real route table — once connected, ask your MCP client to list its tools, or just describe what you want done ("generate a video about our new product launch and post it to Instagram and TikTok," "what's in my review queue right now").

n8n community node

Two real nodes covering the whole API — published on npm as n8n-nodes-getmulticast — current version: v1.4.0 · Released Sept 2026. Install it directly from n8n: Settings → Community Nodes → Install → enter n8n-nodes-getmulticast. Source on GitHub; submitted to n8n for official Verified-node review.

GetMulticast (action node) covers every resource in this API — Video, Post (including Analytics), Comment, Pending Reply, Webhook, RSS Feed, Recycling Rule, Account, Credits. The Video resource (Create/Get/List/Delete) plays the same role Blotato's "Visual" resource plays in their own n8n node: generate an AI video, poll until it's ready, then feed the result into Post > Create — the same shape as Blotato's own "idea in Telegram → AI video → approve → publish everywhere" workflow templates.

GetMulticast Trigger (trigger node) starts a workflow on a real event (post.published, post.failed, comment.received, reply.sent) with full webhook lifecycle management — activating the workflow registers the webhook automatically (POST /webhooks), deactivating it removes the registration (DELETE /webhooks/{id}), and every delivery is HMAC-signature-verified before reaching your workflow. No manual URL-copying step, though n8n's own generic Webhook node still works too if you'd rather register the URL yourself.

Zapier

A Zapier integration isn't listed on the Zapier App Directory yet — that requires Zapier's own developer review process (their own account, their own submission/review cycle — the same shape as any platform app review). In the meantime, Zapier's built-in Webhooks by Zapier action/trigger works against this API directly with no custom app needed: use a "Custom Request" action for any endpoint on this page (set the Authorization header to your API key), and a "Catch Hook" trigger + this API's POST /webhooks to receive events in real time.

What's not here yet

Built honestly, not oversold — here's what v1 deliberately doesn't include yet:

• Analytics for Instagram, Facebook, Threads, LinkedIn, TikTok, and Google Business — see GET /posts/{id}/analytics above for exactly which permission each is missing (a real, specific gap, not a vague "coming soon"). YouTube, X, Pinterest, and Bluesky are real and live today.
• A published, listed Zapier app — the code path above (Webhooks by Zapier) works today without one.

Changelog

Version	Date	Changes
v1.1	Sept 2026	Added AI Video Generation — POST/GET /videos, GET/DELETE /videos/{id} — same engine as the dashboard's Generate Video button. MCP server bumped to v1.3.0 (37 tools).
v1.0	Sept 2026	Initial public release — every endpoint on this page, all real, all live from day one.

Questions or found a bug? support@getmulticast.com