When Feyi Ogunsanya found herself spending more time piecing together logs than building AI applications, she realized the problem wasn’t the models; it was the lack of visibility into what they were doing.
Instead of relying on cloud dashboards, scattered logs, and endless debugging sessions, she built Axon, an open-source observability platform that helps developers understand how AI agents think, reason, and make decisions in real time.
Better yet, it’s completely open source.
“I wanted AI observability to be as simple as running a command and immediately seeing what my AI application was doing.”
— Feyi Ogunsanya, Creator of Axon
What Is Axon?
Axon is a local-first, open-source observability platform designed for AI applications.
Built on OpenTelemetry, it captures prompts, tool calls, execution traces, latency, token usage, and other telemetry, giving developers end-to-end visibility into how AI agents behave during execution.
Unlike many commercial observability tools, Axon integrates with existing AI applications without requiring proprietary SDKs or vendor lock-in, allowing developers to adopt it within their existing workflows.
Since its launch, the project has grown into an actively maintained open-source tool, recording more than 1,000 weekly npm downloads.
Why AI Developers Need Better Observability
As AI applications become increasingly autonomous, debugging them has become significantly more complex.
Unlike traditional software, AI agents don’t always fail with clear error messages. They may retrieve the wrong context, invoke the wrong tool, or produce unexpected responses while appearing to complete successfully.
That leaves developers jumping between logs, terminals, and cloud dashboards trying to answer a simple but critical question:
What actually happened?
For Ogunsanya, this challenge became impossible to ignore.
“I realized I was spending too much time trying to understand AI behavior instead of improving it. I wanted one place where developers could see everything that happened during execution.”
Enter Axon
Axon is a lightweight command-line tool that enables developers to collect, inspect, and visualize AI execution traces locally.
Using OpenTelemetry, it receives traces from AI applications and presents them in an intuitive dashboard, making it easier to inspect prompts, tool calls, execution flow, latency, and model responses—all from a single interface.
Rather than forcing developers into another cloud platform, Axon keeps observability close to where development happens.
Key Features
- Local-first architecture
- Built natively on OpenTelemetry
- No proprietary SDK or vendor lock-in
- Live AI execution tracing
- Prompt and tool-call inspection
- Execution trees and waterfall timelines
- Conversation transcripts
- SQLite-backed local storage
- Compatible with OpenLLMetry and other OTLP-enabled frameworks
“I built Axon to fit naturally into the way developers already work. You shouldn’t need a cloud account just to debug an AI application.”
Growing Adoption
Axon’s growing adoption within the developer community reflects the increasing demand for transparent AI tooling.
The project has surpassed 1,000 weekly npm downloads, with active releases and continuous community contributions, positioning it as an emerging open-source solution for AI observability.
Building Trust in AI Systems
For Ogunsanya, Axon is about more than debugging, it’s about making AI systems easier to understand and ultimately more trustworthy.
“AI is changing how we build software, but developers need better tools to understand and trust these systems. My goal is to make AI observability simple, open, and accessible to everyone building with AI.”
As AI agents continue to power more applications, observability is becoming a foundational part of the development stack. By making these tools open, local-first, and accessible, Axon offers developers a practical way to better understand the systems they’re building.
Learn More
- GitHub: https://github.com/langchain-tracer/Axon
- npm: https://www.npmjs.com/package/@axon-ai/cli
- Documentation: Available in the project’s README.
About the Creator
Feyi Ogunsanya is a software engineer and AI builder focused on creating developer tools that simplify AI development. She is the creator of Axon and welcomes community contributions, collaboration, and feedback on the project.
