Start with the request
TraceFlow keeps the request at the center. Instead of only showing that an API call happened, the goal is to capture useful information about what happened during that request.
TraceFlow was created around a simple idea: when a developer sends a request, they should be able to understand what happened to that request—not just see its response.
TraceFlow keeps the request at the center. Instead of only showing that an API call happened, the goal is to capture useful information about what happened during that request.
As the system grows, a request can cross services and make external calls. TraceFlow is being built to make that path easier to follow inside the same application.
The purpose of tracing is not simply collecting data. It is helping a developer move from a slow or failing request to the part of the request that needs attention.
The next stage is deeper span-level visibility, especially around external requests, so developers can see where time is actually being spent and investigate bottlenecks faster.
Developers often have to move between request tools, logs, metrics, and tracing systems to understand a problem. TraceFlow is being built to bring more of that context into the same workflow: make the API request, capture its trace, inspect the information around it, and use that information to locate the bottleneck faster.
The idea was shaped by working with tracing tools such as Jaeger. Jaeger showed what distributed tracing could do, but TraceFlow was created with a different product experience in mind—one that feels closer to the developer’s request workflow and makes investigating a trace straightforward.
Built as a learning project. Designed as a real system.
TraceFlow’s architecture has been designed from the ground up, including request context isolation, trace handling, storage, and the path from an incoming request to the developer dashboard. Building it also meant digging into how request context can be isolated and propagated with AsyncLocalStorage and how a tracing system should behave as the number of requests grows.
what comes next
The long-term direction is to capture spans automatically, including external requests, so developers do not have to manually instrument every part of their application just to understand where a request is spending its time.
Explore TraceFlow