TattvaDocs
Core concepts

Trackers

Questions that re-run on a schedule, with each run classified and kept, so a one-off answer becomes something you watch.

A tracker is a set of saved questions that re-run on a schedule. Each run is classified, so you can see at a glance which questions need you. Every run is kept, so you can see how an answer changed over time.

Tracker runs use the same metric definitions as Reasoning. A figure in a tracker is computed the same way as the same figure in an answer.

Trackers list

The Trackers page

Sidebar → Trackers (/tracker).

  • Create Tracker (top right) opens the create form.
  • Each row shows the tracker name, N Questions · cadence, Last update, and a Run now button.
  • Click a row to open the tracker.

Create a tracker

Two ways:

  1. From a Reasoning answer. Click Save as tracker in the answer toolbar. Tattva creates the tracker straight away, with your question as its first question, a daily cadence, and not marked as shared. Everyone in the project can still see it. It then opens the new tracker.

  2. From scratch. Click Create Tracker. Fill in:

    • Name
    • First question (optional)
    • Cadence: Hourly, Daily, Weekly, or Manual / on demand
    • Shared with team: on by default. This is a label; everyone in the project sees every tracker.

    Click Create.

You can also start a tracker from a goal. See Goals.

The tracker page

Header actions:

  • Run now runs every question immediately, whatever the cadence.
  • Add Question opens Add a tracked question.
  • Edit renames the tracker.
  • The ⋯ menu has Rename, Pause tracker (stops scheduled runs; later Resume (daily)), and Delete tracker….

Tabs: Questions, History, Delivery.

Questions

The left side lists Tracked Questions. Each card shows the question's status, a one-line summary of the latest run, its cadence, its rule, when it runs next, and an Edit link.

Click a question to see its latest run on the right: what changed, why it matters, and the next best action (or key takeaways). Continue in chat opens the question in Reasoning.

Tracker questions

Question states

StateWhat it means
CriticalThe latest run broke a Relative delta rule.
Needs attentionThe question has an Absolute threshold or Hypothesis invalidated rule. See Rules: these rules flag every run.
All clearA Relative delta rule did not trip, or the question has no rule, or an LLM judge rule.
SkippedThe run could not compute the figure. The card reads Skipped — reason, for example that no metric could be found for the question.
Not run yetThe question has no runs.

A question with no rule shows a no rule chip. Add a rule if you want Tattva to flag changes for you.

A skipped run is not an all-clear

Skipped means Tattva could not work the number out, so it did not guess. Read the reason. It usually points to a metric name the project does not define yet, or a source that is not connected.

Rules

Click Edit on a question to open Question settings. There you can change the question text, the Cadence, and the Classification rule, then click Save schedule.

RuleWhat it does
No ruleNothing is flagged. Runs show All clear.
Relative deltaClassifies each run by the size of the change the run measured. If the change is larger than your critical if Δ > percentage, in either direction, the question is Critical. Otherwise it is All clear. The critical if Δ < value is not used.
Absolute thresholdEvery run is marked Needs attention. The upper and lower values you enter are saved but not checked.
Hypothesis invalidatedEvery run is marked Needs attention.
LLM judgeThe prompt is saved, but runs are not judged by it. They show All clear.

Leave the metric slug blank to use the question's own metric.

Cadences

  • hourly: for fast-moving operational figures.
  • daily: where most trackers live.
  • weekly: for slower KPIs.
  • Paused: use Pause tracker in the ⋯ menu. The tracker then runs only when you click Run now. Trackers created with Manual / on demand behave the same way.

There is no time-of-day setting. The question card shows when the next run is due, for example "next in 37m".

Question settings

History

Every run, newest first, per question. Each row shows when it ran, its classification, the change, and a short summary. Expand a row to read what changed, why it matters and the next best action, or Continue in chat →. When the tracker is linked to a goal, a small chart shows the distance to target over time.

Tracker history

Delivery

Send runs to Email, Slack or Teams. Click Add destination, pick the channel, enter the address or webhook URL, and choose Summarised or Full. Test send runs the tracker and delivers straight away.

A run is delivered when it has a classification and something changed materially since the previous run. Critical and Needs attention runs also create an in-app notification. See Notifications for workspace defaults.

Tips

  • Start with a few trackers. Flags only help if someone reads them.
  • Give every question a rule. Without one, nothing is ever flagged.
  • Read skipped reasons. They tell you what the project still needs to define.
  • Link trackers to goals. A goal-linked tracker also tells you how far you are from the target.
Trackers — Tattva Docs