Traceway unifies every observability signal a team needs — logs, distributed traces, metrics, exception tracking, session replay, and AI observability — into one MIT-licensed platform that runs entirely on your own infrastructure. You point any OpenTelemetry SDK at it over OTLP/HTTP and data starts flowing immediately, with no Collector process, no proprietary SDK, and no per-event pricing. The stack is intentionally minimal: a Go/Gin backend with ClickHouse for telemetry storage and PostgreSQL for relational data, fronted by a SvelteKit dashboard. An embedded SQLite mode with zero external dependencies is also available for smaller deployments or local development.
Where Traceway stands out from alternatives like Datadog or a DIY Prometheus-Loki-Tempo stack is the depth of its AI observability layer. It captures LLM cost, token counts, latency, and full conversation traces across providers, currently with native support for OpenRouter and any OTel-compatible AI gateway. This makes it a natural fit for teams building LLM-powered applications who want production-grade tracing without routing sensitive prompts through a third-party SaaS. Alerts can be routed to Slack, GitHub, email, or webhooks, and Apdex plus Impact-Score ranking surfaces the endpoints that matter most.
Traceway supports backend frameworks across Go (Gin, Chi, Fiber, net/http), Node.js (NestJS, Hono, Cloudflare Workers), PHP (Symfony), and frontend stacks (Next.js, React, Vue, Svelte) with session replay included. Mobile coverage extends to Flutter, Android, and React Native. The project ships 29 releases since launch, maintains an active Discord community, and offers a managed cloud option running the same MIT codebase for teams that prefer not to operate their own infrastructure.
