axonpush
Trace AI agent calls, tools and backend events in one place; investigate failures, replay them in CI and evaluate changes
Last verified:
What is axonpush?
AxonPush is an observability platform for AI agents that provides unified, real-time tracing and debugging across agents, backend services, and OpenTelemetry spans—all joined by a single trace ID. It helps teams understand why their AI agents fail by auto-instrumenting popular frameworks with sub-100ms latency event delivery.
axonpush pricing
Pricing model: Freemium
Not publicly disclosed; requires contacting team or requesting private beta access
axonpush pros
- Unified tracing across agent, backend, and OpenTelemetry spans on one timeline
- Auto-instrumentation for LangGraph, CrewAI, Anthropic, OpenAI, LangChain with minimal setup (one config line if already shipping OpenTelemetry)
- Real-time sub-100ms latency event delivery with hosted MCP service
- Supported in both Python and JavaScript/TypeScript SDKs
- Personal customer onboarding included
axonpush cons
- Pricing not publicly disclosed—requires contacting sales or joining private beta
- Early stage product (v0.0.15 on PyPI indicates pre-release)
- Limited information on free tier, self-hosted, or deployment options
- Few third-party reviews or independent case studies available
Frequently asked questions about axonpush
What AI frameworks does AxonPush integrate with?
LangGraph, CrewAI, Anthropic, OpenAI Agents SDK, LangChain, Deep Agents, and the Python observability stack (Loguru, structlog, OpenTelemetry, Sentry).
Do I need to rewrite my code to use AxonPush?
No. If you're already shipping OpenTelemetry, integration requires just one config line. Otherwise, auto-instrumentation is available for supported frameworks.
What is the latency?
AxonPush publishes, subscribes, traces, and delivers agent events with sub-100ms latency.
How do I get started?
Install the Python SDK via PyPI (pip install axonpush), or contact the team to request private beta access to their hosted platform.