# Plexus > Plexus is the observability system for hardware fleets — built for the teams and the AI agents who share the watch. Connect a device or service with a three-line SDK (Python or TypeScript) or by pasting one prompt into a coding agent; dashboards build themselves from the first data that lands; monitors and alerts watch every connection; and operating the fleet happens in the app's built-in AI. Hardware-native: edge buffering when the link drops, structured schemas you can query, from an ESP32 on a bench to a bus on orbit. Plexus is the operations frontend for the systems a team already runs. Everything is a connection: point Plexus at a device, a service, or an existing telemetry store and it becomes a live, queryable part of the fleet — no data migration, no rip-and-replace. Dashboards are the UI: they generate automatically from real metrics on first ingest and grow with the system. Setup is agent-native — a coding agent pasted one prompt can instrument a codebase, claim a write-only API key the owner approves in one click, and hand back a working dashboard. ## What it does - Connect: three-line SDKs for Python and TypeScript with batching, retry, and edge buffering; a bash bootstrap for Linux devices; or paste-a-prompt setup where a coding agent instruments your code for you. - Observe: dashboards auto-build from the metrics actually flowing — device overviews, service health, fleet views — and stay live. - Alert: event and threshold monitors on any source or connection, with delivery to email and Slack and a full audit trail of every alert's lifecycle. - Operate: the app's built-in AI works the fleet with you — querying telemetry, proposing dashboards, and acting through the same UI you see. - Identifiers: source slugs (e.g. `drone-001`); metrics as `sourceSlug:metricName`; time ranges 1m–30d. ## Product - [Home](https://plexus.company): The observability system for teams and agents — connect, observe, predict. - [Pricing](https://plexus.company/pricing): Usage-based self-serve tier and enterprise managed deployment. - [Deployment](https://plexus.company/deployment): How managed enterprise deployment works; a Plexus engineer connects it to your stack. - [App / Sign up](https://app.plexus.company/sign-up): Sign up and add a card — usage-priced from the first data point. ## Set up by agent - [Coding-agent setup guide](https://plexus.company/agent.md): Step-by-step guide a coding agent (Claude Code, Cursor, etc.) follows to instrument a project — SDK install, one-click API-key claim flow, what to track, dashboard handoff. ## Compare & guides - [Compare Plexus (overview)](https://plexus.company/compare): How Plexus compares to Datadog, Grafana, Splunk, PagerDuty, ClickHouse, InfluxDB, NVSentinel, Memfault, and Foxglove — index of all head-to-heads. - [Datadog alternatives](https://plexus.company/compare/datadog-alternatives): Ranked, honest field of Datadog alternatives — why teams leave (cost, alert fatigue, lock-in) and which tool fits which problem. - [Plexus vs Datadog](https://plexus.company/compare/plexus-vs-datadog): Datadog is cloud-software observability priced by ingest; Plexus is hardware-fleet observability on your existing stack. - [Plexus vs Grafana](https://plexus.company/compare/plexus-vs-grafana): Grafana visualizes what you wire up; Plexus auto-builds the fleet view from the data itself. Complementary layers. - [Plexus vs Splunk](https://plexus.company/compare/plexus-vs-splunk): Splunk indexes everything at volume-based cost; Plexus is fleet observability without the ingest tax. - [Plexus vs PagerDuty](https://plexus.company/compare/plexus-vs-pagerduty): PagerDuty routes alerts to humans; Plexus is where the telemetry, dashboards, and monitors behind those alerts live. - [Plexus vs ClickHouse](https://plexus.company/compare/plexus-vs-clickhouse): ClickHouse stores telemetry; Plexus is the operations frontend that reads from it. Not alternatives. - [Plexus vs InfluxDB](https://plexus.company/compare/plexus-vs-influxdb): InfluxDB is a time-series store; Plexus is the fleet UI and alerting layer above the store. - [Plexus vs NVIDIA NVSentinel](https://plexus.company/compare/plexus-vs-nvsentinel): NVSentinel is single-vendor GPU health; Plexus is cross-vendor whole-fleet observability. - [Plexus vs Memfault](https://plexus.company/compare/plexus-vs-memfault): Memfault watches device firmware; Plexus watches the whole fleet — devices, services, and stores together. - [Plexus vs Foxglove](https://plexus.company/compare/plexus-vs-foxglove): Foxglove visualizes robotics data; Plexus runs the fleet's metrics, dashboards, and alerts. ## Customers - [Customer stories](https://plexus.company/customers): How hardware and infrastructure teams run Plexus in production. ## Resources - [Resources](https://plexus.company/resources): Guides, calculators, and reference material for hardware and infrastructure teams. - [Fleet cost calculator](https://plexus.company/resources/calculator): Estimate observability costs at scale. ## Blog - [Blog](https://plexus.company/blog): Engineering and product posts on observability, telemetry, and fleet reliability. - [Anomaly detection](https://plexus.company/blog/anomaly-detection): Detecting anomalies in device and fleet telemetry. - [Raspberry Pi observer](https://plexus.company/blog/pi-observer): Edge fleet monitoring from a Raspberry Pi (SDK on-ramp). ## Company - [Company](https://plexus.company/company): About Plexus — who we are and why we build this. - [Contact](mailto:info@plexus.company): Talk to the team about enterprise deployment or partnership. ## Documentation - [Docs](https://docs.plexus.company): Full SDK reference, API docs, and integration guides.