Ecosystem

DeepXplore is an AI-powered performance intelligence platform that uses GreptimeDB as its unified telemetry backend — turning metrics, logs, and traces into automated root cause analysis and remediation.

Telemetry flows in through standard protocols — Prometheus Remote Write and OTLP — with no custom integration code. Once data lands, the same store powers threshold validation, dashboards, anomaly detection, and AI root cause analysis simultaneously.

Because GreptimeDB holds metrics, logs, and traces in one engine, DeepXplore's AI agents correlate evidence across signal types seamlessly. The DeepXplore Agent runs in-cluster and pulls RCA tasks outbound — no inbound exposure required. When evidence is strong enough, DeepXplore Code auto-implements fixes, runs tests, and deploys, closing the loop from detection to remediation.

• AI Root Cause Analysis — ranked, evidence-backed root causes in seconds
• DeepXplore Code — AI auto-remediation from root cause to deployed fix
• Performance Testing — realistic traffic simulation and user journey testing
• Anomaly Detection — seasonality-aware ML that predicts SLA breaches before users are impacted

Hebo AI is a conversational AI platform that uses GreptimeDB as the observability backend for its AI Gateway — providing real-time visibility into LLM traces, token usage, inference latency, and Gen-AI signals.

Hebo ingests telemetry through GreptimeDB's native OTLP endpoint, collapsing a typical stack of Redis, ingestion workers, transformation pipelines, and a separate database into a single engine. Traces, metrics, and logs flow directly from SDKs over HTTP with no intermediate infrastructure required.

GreptimeDB's time-series-native storage and object-storage-first architecture make it a natural fit for AI observability workloads. Span attributes from the emerging Gen-AI OpenTelemetry semantic conventions — model names, token counts, input/output messages — are automatically flattened into columns via dynamic schema, with no migrations needed as the conventions evolve.

• AI Gateway — unified access to SOTA LLMs with auto-logged calls and token tracking
• Observability — conversation-level traces, latency metrics, and tool call inspection
• Evals — human-first evaluations written in Markdown, runnable from CLI