Reports
Reports help you move from one translation to a pattern: which languages account for most of the work, which measurement's answers changed, where a rule keeps failing. They live in Radar, on two screens:
- Radar → Overview summarizes the current month: the average score of your measurements, words generated, today's events, and active measurements.
- Radar → Explore charts events and measurement answers over any of its date ranges, and saves those charts as reports for the organization.
Both read the same recorded events and answers as the rest of Radar, so a report always leads back to the translations behind it.
Plan availability
Explore and the overview's report tiles come with the Enterprise plan; without it, they say that reports are not on your plan. Measurements, their answers under Radar → Events, and the overview's Active measurements tile are available on every plan.
Token usage and cost
Detailed token and cost data lives on the Usage page in Settings.
Overview#
| Tile | What it shows |
|---|---|
| Average score | Daily average score for the month, from every measurement's answers |
| Words generated | Words your engines translated this month, compared with the same days of the previous month |
| Events today | Events recorded today |
| Active measurements | Measurements that are on, and what they are reading |
Each day of the score panel pools that day's answers from every measurement: a pass counts 1, a fail 0, and a 0 to 100 score its value divided by 100. N/A answers and structured answers such as edit patterns are left out. The headline averages the days that have a score, without weighting them by answer count, and the line below compares the month so far with the same days of the previous month, in points.
Because the panel pools every measurement, it mixes criteria that point in different directions - a higher edit distance means more text changed. Treat a movement in it as a reason to open Explore with one measurement selected, not as a quality verdict.
Build a report in Explore#
Start a report
Open Radar → Explore and select New report, or open Examples and adjust a report drawn from your data.
Choose the source and data
Under Source, choose Events, Measurements, or both - each data series keeps its own source. Under Data, select Add data and choose what to count or calculate.
Filter and group
Under Filters, select Add filter, choose a field, and use is any of, is none of, or is empty. Under Group by, choose a time grain - Total, Day, Week, or Month - and an optional Breakdown, with Then by for a second one.
Chart and save
Choose a chart and a date range, then select Save report.
The charts are line, bar, number, horizontal bars, pie, heatmap, scatter plot, dot plot, and histogram. The date ranges are the last 7, 30, or 90 days and the last 12 months; each ends today, so today counts while it is in progress. Days, weeks, and months follow the organization's timezone, set under Settings → General, and weeks start on Monday.
Events#
| Data | Calculation |
|---|---|
| Events | Recorded events |
| Strings | Distinct content units among them |
| Input tokens, Output tokens | Tokens used by Lingo.dev localizations |
| Payload properties | A count, distinct count, sum, average, minimum, or maximum of a stored payload field |
Filter by Event type before counting - a localization's start and its completion are two events. Use Words generated on the overview for word volume.
Break events down by Target locale, Source locale, Produced by, Provider, Model, or Engine, or by payload fields such as Project, File, String key, and Repository. A connected GitHub repository records an event for each localization key a commit adds, changes, or removes, so events grouped by Repository show where your codebase changes localized content.
Measurements#
| Data | Calculation |
|---|---|
| Answers | Every answer, N/A included |
| Scored answers | Answers that are not N/A |
| Not applicable | N/A answers |
| Passed answers, Failed answers | Pass-or-fail answers that passed or failed |
| Average score | Normalized scores summed over answers that are not N/A, divided by their count |
| Pass rate | Passed answers divided by pass-or-fail answers that are not N/A |
Select one Measurement as a filter or breakdown before comparing scores. Without it, Average score combines unrelated criteria - and pass-or-fail answers with 0 to 100 scores, unless Answer type narrows it. Measurement reports can also break down by Judge model - the model that answered - and by Score, which sorts answers into bins of ten for a histogram.
Glossary adherence and rule adherence answers show how consistently your glossary terms and rules are applied; edit rate and edit distance show how often and how much people change translations.
Breakdowns and payload keys#
A breakdown keeps the largest groups and combines the rest into Other: the top 10 without a time grain, the top 5 with one, and the top 100 without Other when Then by adds a second breakdown. Records without the field appear as Not set.
Filters and Breakdown also list the keys found in the events' payloads by their path, such as originalEvent.entity.type for a Zendesk article. A report reads one payload key at a time. The list comes from a sample of the organization's payloads, so a key that few events carry can be missing.
Keep and share reports#
A saved report stores its name and definition - series, filters, grouping, chart, and date range - not its data. Opening it reruns the query, so the same report can show different values later. Every member of the organization sees the organization's reports and can rename, overwrite, or delete any of them; a deleted report cannot be restored.
After changing a saved report, select Save to overwrite it, Save as new to keep the original, or Discard. Download CSV exports the table under the chart exactly as displayed, including the Other group.
Averages of days and pooled values#
Averaging daily percentages and pooling answers can give different results. If one day has 1 pass from 1 answer and another 0 passes from 9, their daily rates average to 50%, but the ten answers together pass 10%. The overview headline averages days and reports 50%; an Explore Number chart over the same range pools its answers and reports 10%. Read a pooled value when you need the rate for a whole population.
Where the earlier reports went#
| Earlier report | In Radar |
|---|---|
| Word Generations | Words generated on the overview |
| Top Locales | An Explore Events report with Input tokens or Output tokens, broken down by Target locale or Source locale |
| Change Rate | An Explore Events report broken down by Repository |
| Average Scores | The overview score panel, or Average score in Explore broken down by Measurement |
| Terminology Coverage | Glossary adherence answers in Explore |
| Rule Adherence | Rule adherence answers in Explore |
Reports via MCP and the API#
The Lingo.dev MCP server can query events and measurement answers and create, update, and delete saved reports, so an AI assistant can answer a question about your translation activity from the conversation. Saved reports are also available through the API at /v2/organizations/:orgId/reports.