Publish services to the internet or reach them privately through outbound-only tunnels, with identity, access policies, and connection logs managed at the edge. Use rstream tunnels for local webhooks, private databases and homelabs, LLM and GPU meshes, real-time video, or remote fleet operations with rstream WebTTY. Control these workflows through the CLI, SDKs, APIs, Docker, and Kubernetes.
Today is an important moment for me. rstream is the result of several years of work, and I deliberately waited before launching it here,
“Launch early” is common advice, and often good advice. But rstream carries real traffic and sits on a security boundary. I wanted to trust it with my own traffic before asking anyone else to trust it with theirs. For several months, rstream has carried both my own traffic and partners’ production workloads, and I know it is mature and ready for wider use,
I built rstream so the same connectivity model can follow a service from a laptop into production,
You can start with a simple command:
rstream forward 8080
That gives a local service a public HTTPS endpoint. The same tunnel can be configured in YAML, run in Docker or Kubernetes, or embedded through APIs and SDKs. A service can also remain private, reachable only by authenticated users, applications, or agents under explicit credentials and policies, including through MCP and Agent Skills. Connections originate inside the private environment, so no inbound port is opened and access remains limited to that service,
rstream WebTTY extends the same tunnels from temporary remote terminals to registered fleets, with recordings, live collaboration, and optional end-to-end encryption,
Something important for me: all rstream client code that runs on your machines is open source and inspectable,
Much of the work went into protocol correctness, security, performance, connection lifecycle, and observability. Identity, mTLS, access policies, and connection logs are enforced at the edge,
At launch, rstream runs on a global edge network spanning four regions,
ngrok opened this category and remains an excellent product. As it has evolved into a powerful programmable ingress and API gateway, I took rstream in another direction. Public exposure, private dialing, non-HTTP and real-time traffic, WebTTY, and application-owned connectivity are all first-class parts of the same runtime,
Basic is free. Pro is a flat $49 per project during launch, including 50 simultaneous tunnels and 200 GB, and early users keep that price forever. It remains predictable as usage grows. Enterprise runs on dedicated rstream infrastructure, with unlimited tunnels and bandwidth under contract,
If you try rstream, I would genuinely value your feedback on what you connected, where it got in your way, and what you would need for your application,
Thank you for taking the time to look at something that has mattered to me for a long time,
Hello Product Hunt 👋
Today is an important moment for me. rstream is the result of several years of work, and I deliberately waited before launching it here,
“Launch early” is common advice, and often good advice. But rstream carries real traffic and sits on a security boundary. I wanted to trust it with my own traffic before asking anyone else to trust it with theirs. For several months, rstream has carried both my own traffic and partners’ production workloads, and I know it is mature and ready for wider use,
I built rstream so the same connectivity model can follow a service from a laptop into production,
You can start with a simple command:
rstream forward 8080
That gives a local service a public HTTPS endpoint. The same tunnel can be configured in YAML, run in Docker or Kubernetes, or embedded through APIs and SDKs. A service can also remain private, reachable only by authenticated users, applications, or agents under explicit credentials and policies, including through MCP and Agent Skills. Connections originate inside the private environment, so no inbound port is opened and access remains limited to that service,
rstream WebTTY extends the same tunnels from temporary remote terminals to registered fleets, with recordings, live collaboration, and optional end-to-end encryption,
Something important for me: all rstream client code that runs on your machines is open source and inspectable,
Much of the work went into protocol correctness, security, performance, connection lifecycle, and observability. Identity, mTLS, access policies, and connection logs are enforced at the edge,
At launch, rstream runs on a global edge network spanning four regions,
ngrok opened this category and remains an excellent product. As it has evolved into a powerful programmable ingress and API gateway, I took rstream in another direction. Public exposure, private dialing, non-HTTP and real-time traffic, WebTTY, and application-owned connectivity are all first-class parts of the same runtime,
Basic is free. Pro is a flat $49 per project during launch, including 50 simultaneous tunnels and 200 GB, and early users keep that price forever. It remains predictable as usage grows. Enterprise runs on dedicated rstream infrastructure, with unlimited tunnels and bandwidth under contract,
If you try rstream, I would genuinely value your feedback on what you connected, where it got in your way, and what you would need for your application,
Thank you for taking the time to look at something that has mattered to me for a long time,
Regards / uartnet