Images & workers
Prepare approved agent images and connect the hosts that run them.
Explore this feature →AIrspace / Run agents securely
Run agents inside approved boundaries
Before an AI agent starts work, AIrspace checks the software it will run, who is launching it, and the machine that will host it.
What this means for your team
- Use isolated microVM execution on supported backends with an approved runtime posture.
- Define resource, runtime and network boundaries in the Agent’s approved envelope.
- Apply launch authorization and configured human decisions before scheduling.
- Keep the host, Agent identity, run instance and launching user attributable.
A practical example
A coding assistant launches inside its own controlled machine and may reach only the destinations its configured network policy allows.
How it works
Select a step to read its explanation
- An agent starts from an image: a prepared package containing its operating system and software. AIrspace checks that the image and release are approved for the intended environment, so a user does not accidentally launch unreviewed software.
- AIrspace checks who is launching the agent, which version they selected and whether the relevant policy permits the run. If the customer has configured an approval, the required person must decide before the run proceeds.
- A worker is the physical or virtual machine that runs agents: for example, a cloud VM or a server in your data center. AIrspace selects an enrolled, trusted worker that meets the agent’s requirements, such as available capacity and the required security configuration.
- The agent runs inside its own microVM, a lightweight virtual machine that isolates its work from the host and other runs. Configured limits govern resources and network destinations, so the agent operates within the boundaries chosen for it.
- When the task ends, the run can be stopped through AIrspace. Depending on the configured controls, authorized people can review lifecycle information and recordings to understand what happened.
1 / 5
This example illustrates the journey. Select any step to explore it at your own pace.
Explore more AIrspace features
Model, tool & peer gateways
Control which models, tools and peer agents a run may reach.
Explore this feature →Live sessions & recordings
Follow a live run and review its recording with the right access.
Explore this feature →AI Agent Session Recording Analyzer
Under Development
Investigate recorded agent sessions through evidence-linked findings, timelines and clips.
Explore this feature →Agent threat intelligence
Under Development
Correlate evidence from independent observers to investigate agent behavior.
Explore this feature →