Build AI agents. Promote them. Account for it all.
Build agents on your own knowledge and tools. Promote them Development to QA to Release, version every change, and attribute every token to a business unit.
Or point AETHER at any OpenAI-compatible endpoint, including models you host yourself.
See the model pricesheetBuild the agent. Govern the spend. Deploy the change.
Three screens from inside AETHER, not mockups.
Build
Configure a persona, system instructions, knowledge and tools for an agent, then move it from Development toward Release.
Govern
See spend broken down by business unit, by model and by call type, not just a single total.
Deploy
Design a branching conversation workflow by dragging steps onto a canvas.
Promoted, versioned, attributed, capped.
Anyone can build an agent. AETHER is what lets a platform team say what it cost, who owns it and how it reached production.
A production change goes through a gate
Agents move Development to QA to Release, and the promotion control sits in the builder next to the agent it protects: the same discipline a code deploy already gets.
A live agent cannot be edited by accident
Changing a released agent means a new major version, created with an explicit New Major Version action, never an edit saved silently over what is already running.
Spend that means something to finance
Every token is attributed to a business unit, a model and a call type, and the breakdown exports to CSV, Excel or PDF for whichever report is already due this week.
No agent can quietly run up a bill
Per-agent and per-developer budgets are enforced as hard stops on the platform, not alerts that arrive by email after the spend already happened.
Isolated, encrypted, deployable in your own cloud.
Security is usually the first question asked about an enterprise AI platform. Here is what governs where your data goes, who can see it, and where AETHER itself can run.
Deploy inside your own AWS, Azure or GCP tenant
AETHER does not have to run in a shared environment. It can be deployed inside an AWS, Azure or GCP account your organisation already controls, so the platform runs on infrastructure you operate rather than only ours.
Your data does not train anyone's models
Content sent through AETHER, yours or your agents', is not used to train AI models, ours or any model provider's.
Isolated and encrypted by default
Every organisation's data is kept separate, so one tenant cannot read another's. Traffic is encrypted in transit, and data at rest is encrypted by our hosting provider.
PII Guard
Every message is scanned on the platform before it reaches a model provider. Personal information is masked, removed or refused, based on a policy set per agent, and applied to typed text, screenshots and whatever an agent reads for itself.
Secrets Guard
API keys, access tokens and connection-string passwords get their own policy. A live secret is refused outright; one an agent reads from a file is stripped instead, so the task carries on.
One agent. Six ways to reach it.
Build it once on AETHER, then meet people where they already are: in chat, in Teams, in the editor, or behind your own API.
Published agent page
Every agent gets its own hosted page at a stable URL, its name, description and knowledge sources on view. Open it directly or share the link, no embed code or app install required.
Voice
Give an agent a voice. People speak to it naturally and hear a spoken response in real time, for hands-free interaction and a conversation a chat window cannot offer.
Microsoft Teams
Deploy an agent straight into Microsoft Teams through the Bot Framework, so people reach it inside the tool they already use every day, with no separate app to open.
Web embed
Add a chat widget to any page with a short script tag. Its appearance, position and starting behaviour are configured from the same admin console that manages the agent behind it.
VS Code, via AETHER Coder
Bring an agent into the editor as a full coding assistant: planning, editing files and running commands, grounded in your knowledge base and MCP tools, governed by the same platform as every other surface.
REST API
Integrate with a minimal set of REST endpoints to create agents, send messages, receive responses and manage conversations, built for a developer to get productive in minutes.
From idea to production, in three moves.
Connect your knowledge and tools
Point the agent at documents, databases and MCP tools. AETHER indexes what it needs for retrieval and calls the rest on demand.
Build and test the agent
Write system instructions, choose a model, and try it in the built-in chat before anyone outside your team sees it.
Promote it and watch the cost
Move the agent from Development to QA to Release, and see spend broken down by business unit, model and call type as it runs.
Start building on AETHER.
Create an agent on your own knowledge and tools, test it, then promote it when it is ready. Talk to us first if you would rather walk through it with someone.