API reference
Everything the dashboards do, your own systems can do. It is the same API the product is built on: there is no second, lesser API for customers.

- Base URL:
https://app.reelwire.io. The examples in this manual were run against a development copy athttp://localhost:4000; the paths are the same, only the host differs. - Authentication: a bearer token on every request. See Authentication.
- Limits: every request carrying a key or a session counts against that one credential, 600 a
minute whichever route it calls, and a few routes, signing in among them, have a tighter limit of
their own. A
429says how long to wait. See Rate limits. - Format: JSON in, JSON out, UTF-8. Failures are problem envelopes.
- Versioning: every path is under
/v1/. - Request id: every response carries
x-request-id. Log it. Send your own ULID in that header and Reelwire uses it, so your logs and ours share one id.
Your first request
Who am I: the one call every key may make, whatever its permissions.
curl http://localhost:4000/v1/whoami \
-H "Authorization: Bearer rw_your_key_here"
{
"accountId": "01M1VFRWE3A35H892VAAEKKDA6",
"accountName": "Enterprise 2",
"scopes": ["ingest:write", "posts:read", "media:write"],
"permissions": [
{ "scope": "ingest:write", "label": "Custom feeds - Send raw data: write" },
{ "scope": "posts:read", "label": "Posts: read" },
{ "scope": "media:write", "label": "Media: write" }
],
"kind": "apikey",
"resource": null
}
permissions names what the key may do in the words the key picker and every refusal use. Which
version of Reelwire answered is GET /v1/version, which needs no key.
Sending an event
The most common integration. Your system pushes something that happened; Reelwire draws it into a clip for every channel subscribed to that stream and hands each post over the ways its channel has switched on: by email, by webhook, or through the channel's platform connector, which posts it itself.
curl http://localhost:4000/v1/stream \
-H "Authorization: Bearer rw_your_key_here" \
-H "Content-Type: application/json" \
-d '{
"streamId": "st_30472168d941",
"eventId": "01M2SQMKK3G1NVFXV31DCBVB92",
"occurredAt": "2026-09-18T07:55:00Z",
"locale": "en",
"data": {
"clipData": {
"headline": "Bund yields close at a three-week high",
"subline": "The ten-year settled at 2.41 per cent"
},
"metaData": { "title": "Bund yields close at a three-week high" }
}
}'
Four things to know about that body:
eventIdis a ULID and is the idempotency key. Sending the same one twice does not make a second post, which is what makes a retry safe.- What belongs in
datais whatever the stream's template accepts, published as a schema at/v1/formats/template/{templateId}. Send that shape and it renders; send something else and you get a refusal naming the field. A stream with a field map takes your own shape instead and maps it onto the template's. occurredAtis when the fact happened, not when you sent it.- A picture is a file from your media library, written
media://and its id or name. A web address is refused.
Before sending anything real, POST /v1/stream/validate takes the same body and answers whether it
would have worked, without creating anything.
The whole worked example, with the real responses, is on Sending events.
Where to go next
| If you are | Read |
|---|---|
| Pushing data in from a producer | Sending events, then Streams and feeds |
| Receiving finished posts, and reporting back what became of them | Webhooks |
| Handing your posts on to your own participants (Business, where included) | Syndication |
| Managing colleagues, or reading the plan | Team and plan |
| Watching what happened, and asking for an update, a resend or a delete | Posts and jobs |
| Reading how your posts did on the platforms (Individual and Business) | Analytics |
| Proving what went out and who approved it, for an audit (Business, where included) | Compliance |
| Holding every channel at once, or for a planned period | Channels: Stop everything |
| Scheduling posts somebody wrote | Post lists |
| Setting a workspace up by machine | Brands, Channels, Media library |
| Rendering a clip on its own | Renders and direct export |
| Generating a client | Templates and formats |
| Looking for one route | Every endpoint |
| Choosing what a key may do | Permissions |
| Making a key, showing one again, or checking a rate limit | Authentication |
| Handling a refusal | Errors |
| Reading a long list page by page | Lists and paging |
Two conventions worth knowing up front
A mutation is a POST to a verb, not a DELETE. Removing a brand is POST /v1/brands/{id}/delete. This keeps every action addressable by a permission prefix and keeps
bodies available on actions that need one, such as the moveTo a brand deletion takes.
Lists come in two kinds. Logs are paged by the server; configuration comes back whole. See Lists and paging.