Orama's cloud
Design, train, and ship AI on your own data
Orama's cloud, Lake, keeps your data, trains models and agents on it, and runs what you ship, in workspaces of your own. Ask Orama answers about all of it in plain words.
What it does
From your data to something people use
You build an AI solution in an AI project. Its pipeline is the workflow from your data to a result, and every step happens in the same place.
- 01
Data
Where your data lives and is found.
- Repositories keep files and their whole history
- Drives keep files and folders, with uploads, previews, and history
- Datasets load from a file or a database you connect
- Transcripts turn meetings, calls, and voice memos into searchable text
- 02
Design
Deciding what AI to build, and how.
- A project names the business question and the decision it should improve
- Start from a ready-made AI solution, an opportunity found in your data, or a blank project
- The pipeline lays out each step, and runs only after you approve it
- 03
Train
Training models and agents, and measuring them honestly.
- Every run measures its model on rows held out from training
- Experiments try many settings and promote the best
- Agents answer when you ask, and the ones with a trigger run on their own
- 04
Ship
Putting what you built to work.
- Publish a model and call it from your own code or an AI agent
- Publish search indexes over your datasets
- Run apps at their own address, behind Lake sign-in by default
- Run databases and streams as services
Ask Orama
Ask about your work in plain words
Ask Orama is where you land and the assistant on every page. Ask it about your repositories, projects, data, runs, models, services, and agents.
- It answers from what you are allowed to read, and works as you
- Every change it wants to make is a proposal, and nothing happens until you confirm it
- It hands work to the agent that does the job better: pipelines, apps, experiments
Built to be trusted
Your workspaces, your people, your automation
Workspaces of your own
Each workspace is an isolated environment with its own engine. What it holds is sealed with its own key and kept apart from every other customer's.
People and access
Organizations, teams, and grants decide who reaches what. Members can sign in through your identity provider and stay in step through SCIM, and Activity keeps the history.
REST and MCP
Every capability is a REST call and an MCP tool, with the same permissions and the same audit. An AI agent acts as you, and can never do more than the key you give it.
Who it is for
For teams that build AI on their own data
Data scientists and engineers
One place for the data, the pipeline, the runs, and the model you publish, with a held-out score on every run.
Teams shipping AI to others
Models, search indexes, and apps that the people and systems around you can call.
Organizations with rules to keep
Isolated workspaces, access by grant, single sign-on, and a version that runs on your own infrastructure.
How it fits
Zetta, Lake, and Metis
Three names for three different things, each with its own job.
The engine: Zetta
Orama's data and AI engine. Every Lake workspace runs on an engine of its own.
See the platformThe cloud you work in: Lake
Where your own team designs, trains, and ships, priced by what you use.
The enterprise program: Metis
The Enterprise Intelligence OS: a program our forward-deployed engineers run inside your organization, priced for each enterprise.
Metis pricingTwo ways to run it
Hosted by Orama, or on your infrastructure
Hosted by Orama
In a shared region, pay only for what you use, at one published rate for each kind of usage, with no plans and no monthly fee. A private warehouse or a dedicated cluster of your own comes with a monthly fee that includes usage credit.
On your own infrastructure
On-prem runs on any cloud provider or on bare metal, in any region, with full service (Metis), for a monthly fee that includes usage credit.
See it on your own data
Create an account and start on your own data, or tell us what you want to design, train, and ship and we will show you how it runs on Lake.