Aster Expose
Connect an existing HTTP service to an Aster peer, or publish it through a public HTTPS edge. The edge has its own infrastructure and TLS requirements.
Enterprise peer-to-peer application SDK
Connect services
across machines.
Your service might run in a cloud instance, a customer network, or on the machine under your desk. Aster gives you a way to call it by its cryptographic identity.
Define a typed service, connect to a peer, and control which calls it may make. One shared Rust engine, with APIs for different languages.
Connection path
A service and a call
The interface lives in your code.
Aster carries the call between peers. Your application defines the types, implements the handler, and decides who can use it.
Service on machine B service.py
@service(name="Hello", version=1)
class HelloService:
@rpc()
async def greet(self, req: HelloRequest) -> HelloResponse:
return HelloResponse(message=f"Hello, {req.name}!")Call from machine A shell
$ aster call <peer-address> Hello.greet '{"name":"world"}' { "message": "Hello, world!"}Imports, types, and server setup are omitted here. Follow the complete quickstart for runnable code.
Identity and access
Know which peer
you’re talking to.
A network address tells you where to send traffic. Aster’s peer identity tells you which endpoint is on the other side of the connection.
Read the trust model- 01 Keep an identity across network changes
- Persist the endpoint key to keep the same identity when its IP address changes. Protect and back up that key.
- 02 Authenticate the connection
- Iroh establishes an encrypted connection and authenticates the remote peer. Direct and relay paths use the same peer identity.
- 03 Authorize the work
- Configure admission and method permissions for your application. Knowing a peer’s identity is the starting point for deciding what it may do.
Put it to work
Start with something you can run.
Follow one service from its first call through streaming and access control, or connect an existing service to an AI tool.
MissionControl
Build a DevOps control plane. Collect agent logs and metrics, issue remote commands, and scope operator access.
Remote tools with MCP
Make an Aster service available to an AI application through the CLI’s MCP integration.
Next release
More ways to use the same foundation.
We’re preparing guides for existing HTTP services, durable coordination, and a second agent example alongside the expanded language APIs.
Coordination
Maintain one authoritative owner or active version across a small group of persistent nodes. Use leases and fencing where conflicting work would be unsafe.
AgentXchange Proposed example
A daemon on each machine receives durable messages and triggers locally configured AI agents. The example is planned; it is not available to run yet.
Language surfaces
Choose the API that fits your application.
Rust is the first-class facade over Aster’s shared engine. Python, TypeScript, and the FFI bindings expose that foundation in their own runtime’s idioms.
Rust
Python
TypeScript
Go
Java
Kotlin
.NETThe next release will document each language’s installation, supported features, and platforms. The current documentation covers the published entry points.
Try a call between two machines.
Start with the quickstart, then add the pieces your application needs.