Your GPUs are idle three hours in four. Fill them with the batch work already waiting.
GPUs bought for the working day’s peak sit idle the rest of the week, and using those hours takes a batch platform: a queue, a scheduler that knows when the machines are free, retries, deadlines, somewhere to keep the results. Otium is that platform, on your hardware.
Why the night shift stays empty
-
Sized for the peak
The fleet was bought for daytime load, so most of its hours have nothing scheduled.
-
Batch requests with nowhere to go
Teams with work that could run overnight have no way to hand it to the GPUs.
-
A platform nobody has time to build
A queue, scheduling around daytime use, retries and result storage add up to a project of its own.
What Otium brings to your GPUs
A batch API your teams already speak
OpenAI-compatible batch files and the same SDK, so researchers and engineers submit work without learning a new system.
What batch inference is →Work only in the hours you give it
Nights and weekends, or whatever window you set; work waits for the next window rather than crowding the day.
The two ways to run it →Deadlines and retries
Each batch carries a finish time, and a request that fails or a machine that restarts is retried without losing work.
Deadlines explained →Pipelines with your own tools
Multi-step workflows whose tool steps call systems on your network.
How workflows run →
What you would otherwise build
-
A batch API your code already speaks
OpenAI-compatible batch files, the same SDK and a base-URL change, plus a command-line tool.
-
A scheduler that knows your hours
Work runs only in the hours you give the machines, nights and weekends, and waits for the next window rather than crowding the day.
-
Deadlines, not a place in a queue
Each batch carries a finish time, and the scheduler plans the hours ahead around it.
-
Retries that lose nothing
A request that fails, or a machine that restarts mid-batch, is retried; each job is answered once.
-
Pipelines, not just prompts
Multi-step workflows with tool calls to your own systems and steps a person signs off.
-
Somewhere to see it all
Every batch, run and result in one place, with usage by account and model.
Self-hosted workers
BetaYour GPUs run the work; the platform runs in our cloud: the API, the scheduler, storage and the run views.
$49 a month
Includes 2 billion tokens a month, then $0.02 per million tokens, input and output alike, whatever the model.
First 30 days free, no credit card needed.
On-prem
In developmentThe whole platform, installed and run in your own data center or cloud account, with your GPUs. Nothing runs on ours.
Priced by contract
An annual license with support, set with you for the GPUs you run it on.
Questions people ask first
- Will it get in the way of daytime use?
- Batch work runs only inside the hours you give it, so the day stays yours.
- Does our data leave our network?
- With self-hosted workers, prompts and results pass through Otium’s cloud, encrypted at rest. With on-prem, the whole platform runs in your environment and nothing leaves.
- What does it cost?
- Self-hosted workers are $49 a month including 2 billion tokens, then $0.02 per million tokens. First 30 days free, no credit card needed. On-prem is an annual license set by contract.
- Can we start now?
- Self-hosted workers, yes: create an account and enroll a machine from the portal. First 30 days free, no credit card needed. On-prem is being built with design partners: a pilot on your hardware, on your work.
Put your idle hours to work.
Tell us about your GPUs and the batch work waiting for them. We are looking for design partners to build this with.