Push to an app's registered devices
No company's docs show a full fit yet. The listings below do the job with a limit, or not at all.
What counts as done
One request per batch reaches iOS and Android tokens, the response reports per-token success or a coded failure such as unregistered, and the agent can drop dead tokens from that code alone.
How agents ask for it: send a push to everyone on iOS
, notify these device tokens
, push to the user's phones
Does the job, with a limit 9
It does the job, but something in the docs or a user's account will get in an agent's way. The limit is named in each row.
| Listing | Why |
|---|---|
| Amazon SNS BB Amazon Web Services, Inc. | Platform endpoints wrap APNs and FCM tokens, and SNS disables an endpoint the provider rejects, but a direct publish goes to one endpoint ARN; batches go through a topic with no per-token result in the response. source |
| Courier B trycourier.com, Inc. | Push goes to tokens synced to the user profile through FCM or APNs, but the send returns only a requestId; per-token results come later from message logs and webhooks. source |
| Expo Push Notifications B 650 Industries, Inc. | One POST takes up to 100 messages for iOS and Android, each ticket in order with status and DeviceNotRegistered on dead tokens. acc-25 got ok receipts while nothing showed on iOS, and there is no idempotency key. source |
| Firebase Cloud Messaging BB Google LLC | Admin SDK sendEachForMulticast takes up to 500 FIDs or tokens with per-item responses in order, and UNREGISTERED (404) marks dead tokens; HTTP v1 itself is one message per request. acc-21 got NotRegistered for a live FID and acc-24 an undefined error code. source |
| Knock A Knock Labs, Inc. | Push goes to tokens stored as channel data through your APNs or FCM provider, but the trigger returns only a run id; per-token outcomes come later from message statuses and delivery logs, not the response. source |
| MagicBell C MagicBell, Inc. | Saved APNs, FCM and Expo tokens per user are reached by one broadcast, but the broadcast status gives only summary counts and error messages, not per-token codes. source |
| Novu B Noti-Fire Apps Ltd. | Push steps send to subscriber device tokens through FCM, APNs, Expo and others, but the trigger response carries a transactionId only; per-token outcomes come from messages and push activity tracking afterwards. source |
| OneSignal BB OneSignal, Inc. | One call reaches up to 20,000 subscription ids on iOS and Android, and the response lists invalid_aliases and invalid_player_ids for unsubscribed targets, but not a per-token code such as unregistered versus invalid payload. source |
| SuprSend BB SuprStack Inc | Push goes to user push tokens through connected FCM and APNs vendors, and FCM errors are documented, but the trigger returns message ids, not per-token results; outcomes come from message logs. source |
Doesn't do this job 3
Listings in the category that do something else, with the reason.
| Listing | Why |
|---|---|
| ntfy C ntfy LLC | Reaches only devices subscribed to a topic in the ntfy app, not an app's own APNs or FCM tokens, and returns no per-device result. source |
| Pingram C Pingram | Mobile Push discontinued on 7 August 2026, though pushTokens still appear in the send schema. source |
| Pushover C Pushover, LLC | Reaches devices running the Pushover app by user or group key, not an app's own APNs or FCM tokens. source |
What stops agents finishing
- only an app with registered tokens can be reached
- token format migrations (FCM fid) reject valid devices
Around it
Before: code.app-registers-token. After: notifications.check-delivery.
Where this comes from
Weight 10 of 100 in Notifications: 4 accounts (acc-21, acc-22, acc-24, acc-25) on FCM and Expo, all developers
The same as data: /api/v1/use-cases/notifications.push-to-devices.json, or the MCP tool get_use_case with id notifications.push-to-devices. Other jobs in this category: /api/v1/niches?category=notifications.
For companies
Do agents find, use and choose your tools?
An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.