Inferyes
The OVERBRING Labs Superintelligence API Gateway. We run the models ourselves, on our own premises, and the products we build and the partners we build them with reach them through this one gateway.
- Requests served
- 0
- Tokens generated
- 0
- Documents extracted
- 0
- Text extracted
- 0.0 MB
- Inference time
- 0 s
- Backends online
- 0/0
Traces
Cumulative over the last 7 days, which is as long as the gateway has been counting. Each figure is the lifetime total at the moment it was measured.
0 metered model calls
Models available
0
distinct models across 0/0 backends online.
Latency
- Median
- —
- 95th
- —
across 0 metered call(s) since this gateway went live
Counted from the gateway's own meters, across every workspace it serves.
What it does
One gateway for the things our products keep needing, so that need does not turn into five integrations and a bill that grows with every call. Send a document, a message, or an image for the operations that judge rather than write; get back text, a label, structured JSON, or a verdict.
Documents, messages and images
Send a URL or an upload and you get the text. Send text, or send images to the operations that judge rather than write. What comes back is what the job needs: a label, a route, a summary, a title, tags, structured JSON, sentiment, quality, personal data, a moderation verdict.
Pipelines, not glue code
A preset tree is a graph of model steps that run per document. We build it with you, so your application calls one endpoint instead of orchestrating five.
The models, when you want them
An OpenAI-compatible surface for chat — including images, on the models that read them — embeddings and rerank, plus a runs API with streaming, cancellation and a single acknowledged fetch of the result.
Where a document goes
Text you already hold, and images, skip the first two steps: they go straight to the work.
-
01
Your application sends a URL, an upload, or a piece of text to one endpoint.
-
02
Workers on our machines fetch it and turn it into text, then delete the copy they parked as soon as we have taken it.
-
03
The steps you and we configured run against that text, on the model each step asks for.
-
04
You collect the result once. Taking it is what clears it.
What we keep
Nothing of yours. The document, the text and the answer are held only while a run needs them, delivered once, and then cleared. How that works.
Your numbers, kept separately. Model, task, tokens and latency, per workspace and per token, so you can bill what you use and nobody else's traffic touches yours.
No pricing page, no signup, no free tier. This is not a public service. Access is by invitation, and capacity is shared with our own work.
Who it is for
Our own products run on this gateway, and so do the projects we incubate or build with partners. breek.gr ships its document features on it today.
If you are building something that needs this, and we are already working together, the fastest route is the person you talk to. Otherwise, start a conversation.
Talk to us
Tell us what your product needs to do with a document. If it fits, you will hear what it costs to run and when it can start.
Start a conversation