Runway Media Router and sume/auto: what Sume's router does
Looking for a Runway Media Router equivalent? Sume's video router is sume/auto: it picks the model, never says which ran, and clips run 3 to 10 seconds.
Sume's answer to a media router is sume/auto: send it as the model on POST /v1/videos and Sume picks the model family for you. It does not tell you which family ran, and its documented clip range is 3 to 10 seconds. This page covers only Sume's router; Runway's own page was not read for it, so it says nothing about how Runway's Media Router works or what it costs.
What is sume/auto?
The Video generation docs call it a Sume-only addition. Where you would send a catalog id such as seedance-2.5, you send sume/auto instead when you do not want to pin a family. The same field also takes a pinned id, so one endpoint covers both styles. Auto is the recommended path for new integrations; the older Video 1.0 URL is a compatibility alias for the same pipe.
Over MCP, generate_video routes to sume/auto when you omit payload.model, so an agent gets the router by default and pins a family only when the user names one. A repeated request with the same normalized input and catalog version prices and routes identically, which is what makes an idempotent retry safe.
What does a router hide, and what stays visible?
The main tradeoff is visibility. The table compares the two ways to send a request to POST /v1/videos.
| sume/auto | Pinned id (for example seedance-2.5) | |
|---|---|---|
| Who chooses the model | Sume | You |
| Model named in the poll response | sume/auto, never the family | The id you sent |
| Documented clip length | 3 to 10 seconds | Per model: seedance-2.5 takes 4 to 30 seconds |
| Reserved on submit | Provider list × 1.25 | Provider list × 1.25 |
Can I find out which model ran?
No. The docs say Sume does not disclose which family served the request and that you should not build on any observable trait of the output to infer it. If you need to know the model, for a receipt or a style match, pin a catalog id. The reasoning is in which model did my AI video use.
Does routing change what I pay?
The docs describe billing for /v1/videos as a workspace USD balance, reserved on submit at provider list times 1.25, for every model. They do not describe a separate router fee, so treat the router as free of its own charge only as far as those docs go and check the catalog's pricing_skus for a pinned model.
When should I pin instead?
Pin when you need a clip longer than 10 seconds, a particular model's look, or a repeatable model for an approval flow. The call shape is in video model router API, and the decision itself is in sume/auto or a pinned video model.
Sources
Related posts
More in Models
- Does generate_audio change Seedance 2.0's price?
No. fal says Seedance 2.0 audio costs the same on or off, and Sume's Seedance price has no audio input: seedance-2 reserves about $0.38 per 720p 16:9 second.
- Seedance 2.0 max length: 15 seconds, and how to go longer
Seedance 2.0 clips run 4 to 15 seconds on Sume and on fal. For 16 to 30 seconds use seedance-2.5 ($17.34 at 720p) or chain clips with first_frame.
- What is Seedance 2.0 Mini, and where can you run it?
Seedance 2.0 Mini is a Seedance 2.0 model that CapCut and fal list. On Sume it is the id seedance-2-mini, 4–15 s clips, about $0.19 per 720p second.
- Does a reference video change Seedance 2.0's price?
Yes. A reference video adds 15 assumed input seconds at 0.6× on Sume, so a 10 s 720p clip on seedance-2 reserves $5.67, not $3.78.
Written by Sume