Platform
SSE streaming
Server-sent events stream a chat reply token by token over one HTTP connection, so users see text as it is generated instead of waiting for the whole completion.
The OpenAI convention sends data: chunks terminated by data: [DONE].
On eroq
Set stream: true on /v1/chat/completions; credits are flat per completion, streaming costs the same.
Questions
Does streaming cost more?
No — same flat credits per completion.
Which eroq endpoints stream?
Only /v1/chat/completions, with stream set to true on eroq RP+ and eroq RP mini. Images and speech return the finished file, and video is asynchronous — you poll the job or take a webhook instead of streaming it.
What do the chunks look like?
The OpenAI convention — data: lines carrying a JSON delta, closed by data: [DONE]. Existing OpenAI client code parses it unchanged; only the base URL and the model id differ.