Start a translation run
Start a Glossa translation run over the project's pending strings.
Start a Glossa translation run over the project's pending strings. Answers 202 with the batch; poll GET /v1/batches/{batch} until its status is final. One batch per project runs at a time: while one is active this answers 409 active-batch with its id. When nothing is pending, the batch comes back already completed with total 0. A translation you have edited or reviewed is never overwritten by AI; such pairs are counted in skipped.
deployTokenAuthorizationBearer <token>A project deploy token, minted in the console under Settings, Developer, API tokens. Send the whole value, including the numeric prefix before the |. A token carries a set of abilities (push, pull, manage, read, write) and each operation requires one of them. Tokens do not inherit from each other: a manage token cannot read the glossary unless it also has read.
application/json- body
scope*stringmissing and all: every pair with no value, or whose value was written against an older source. stale: only pairs that have a value written against an older source. selected: like missing, limited to key_ids.
"missing""selected""stale""all"key_ids?array<integer>Limit the run to these key ids. Honoured with any scope.
items <= 2000locales?array<>Limit the run to these target locales. Omit for every target locale.
items <= 200The run was accepted.
application/json- response
A translation run.
id?integerstatus?stringpending and processing are active; the other four are final. Only one batch per project is active at a time.
"pending""processing""completed""completed_with_errors""failed""cancelled"trigger?stringWhat started the run.
"manual_missing""manual_selected""auto_edit""language_added""retry_failed""push"total?integerKey and locale pairs sent for translation.
completed?integerPairs the run finished processing.
failed?integerPairs that failed.
skipped?integerPairs the run covered and will not write. A translation you have edited or reviewed is never overwritten by AI; such pairs are counted here.
skipped_reasons?|Skipped pairs by reason. human_value_stands: a person's translation is in place. human_value_stale: a person's translation is in place and the source has changed since, so only a person can update it. source_moved: the source changed while the run was in progress; running again fixes it.
tm_served?integerPairs answered from translation memory. They are not part of total.
language_breakdown?|Pairs per target locale.
started_at?|date-timefinished_at?|date-timecurl -X POST 'https://api.transglot.ai/v1/translate' \ -H 'Content-Type: application/json' \ -d '{ "scope": "missing", "locales": [ "fr", "de" ]}'{ "id": 0, "status": "pending", "trigger": "manual_missing", "total": 0, "completed": 0, "failed": 0, "skipped": 0, "skipped_reasons": { "property1": 0, "property2": 0 }, "tm_served": 0, "language_breakdown": { "property1": 0, "property2": 0 }, "started_at": "2019-08-24T14:15:22Z", "finished_at": "2019-08-24T14:15:22Z"}