Docs

Every ingestion path, written as the configuration that works plus the two or three things that catch people out. All of it applies on the free plan — 5 tags, €0, no card, except Excel reports and read-only chart links, which are on Pro and Business.

Getting data in

Pick the one you already run. Every way in is on every plan, the free one included — what the plans change is how much you can send and keep.

Home Assistant

The Tag Historian integration, installed through HACS: pick entities in a dialog with your tag quota in view, import the history the recorder already holds, keep a tag through an entity rename, and get a Repairs card when something goes wrong.

Home Assistant without HACS

Home Assistant's own InfluxDB integration, pointed at Tag Historian: a dialog under Devices & services, a short YAML filter, and the four settings that decide whether it works.

Leave InfluxDB, keep the history

Keep your InfluxDB, or drop the server without losing the years you have: a resumable copy of the history, the switch of live data, and a Grafana dashboard by entity.

Telegraf

Point a stock Telegraf outputs.influxdb_v2 at Tag Historian, and prune the inputs first.

MQTT

Run a collector next to your broker — Mosquitto, zigbee2mqtt, Tasmota. Topic mapping, payload shapes, and what happens at the tag quota.

OPC UA (beta)

Subscribe to PLC tags from a collector inside the plant network. Security policies, quality codes, and no inbound firewall rule.

Sparkplug B (beta)

Subscribe to a Sparkplug B namespace on your own MQTT broker. One edge node can announce hundreds of metrics, so: filters, a tag budget, the full list of what it announced before you collect any of it, and nothing published but a rebirth request, plus host STATE if you set HostId.

Modbus TCP (beta)

Poll energy meters, older PLCs and RTU gateways from the same collector. 40001 is Address 0, word order, scaling, a CSV tag list, and what happens when a device stops answering.

Looking at it

Stored data you cannot see is a backup, not a historian. There is a chart in the app; this is for the dashboard you already have on the wall, the AI assistant you ask instead of opening a dashboard, and the spreadsheet or the read-only chart you send to someone who has neither.

Being told

Nobody watches a chart at three in the morning. When a value crosses a limit or a source goes quiet, the alert has to find a person — by email, on a phone, or in the chat they already read.

Reference

The endpoint itself, and the three questions any serious evaluator asks before sending real data anywhere: how long it is kept, where it lives, and how it is protected.

Deciding whether to use this at all

Written to be useful even if the answer is no. Two of these three pages will tell some readers they do not need us.

Something missing?

These pages come from documentation written for the product rather than for search, so they cover what exists and stop where it stops. If a page is wrong, out of date, or silent about the thing you needed, say so — a correction is worth more to us than a signup.

support@taghistorian.com