ai Q3 2026
AI observability
Monitoring for AI and LLM workloads: token usage, model latency, prompt and completion pairs, and cost per inference across providers.
What's shipped, what's in progress, and what comes next.
Being built right now.
Monitoring for AI and LLM workloads: token usage, model latency, prompt and completion pairs, and cost per inference across providers.
Committed and queued up.
Simplified self-hosting with a single-command deployment and automated configuration.
On the radar, still being scoped.
Dedicated dashboards for API endpoints with per-route latency, error rates, request volume, and payload size tracking.
Add the SDK, point your OTLP exporter at Maple, and traces arrive.
maple.dev: observability on OpenTelemetry