Loading translation workspace

Preparing projects and storefront languages.

Initial requests are still running.

Initial data load failed. Reload the data before running sync, batch, review, or write actions.

Full-Cycle Automation

On Paused Loading

Daily store sync, translation batches and safe writes. Review items remain manual.

Automation state: Enabled Paused Loading Current run: Loading # active None

Automation is paused, but run # is still active.

Worker is not responding. Run # has work ready for processing, but no step was claimed for s.

Run # :

Workflow steps processed: / (known so far)

Last worker activity: (s ago)

Steps Completed Running Waiting Queued Failed Next check: (s)

Standard batches in this run (all languages) Total Pending Submitted Completed Failed

Language	Lane	Current stage	Status	Next action
{{scope_label}}	{{lane}}	{{phase_label}} {{error_text}} Batch #{{batch_reference}} {{reason_label}}	{{status_label}}	{{available_at_local \|\| '-'}} {{timezone}}

Automatic work is complete. Manual Review remains: .

Last run ended with errors:

Automation settings could not be loaded for the selected project.

Timezone

Daily sync

Standard

SEO lane

Media Alt lane

Alt source method

Batch items

Batches / cycle

Submitted items / day

Failure limit

Auto write

Next sync: Last success:

Project

Storefront

Target language

Hidden store languages can be translated before they are visible in CS-Cart. “Jobs paused” only pauses this tool; it does not publish the language.

Entity

Field

Status

Company

Limit

Selection changed. Tables still use the previously applied filters until Apply is pressed.

AI usage and cost

Project total from the mapped OpenAI cost export, with attributed detail by task, entity, language and model. Every event keeps the tariff version effective when it occurred.

Project API credential

Use the same secret as AI_USAGE_PROJECT_TOKEN in the project server configuration. Saving an existing credential code rotates its token. TranslationHub stores only its SHA-256 hash and never displays the token again.

Credential saved. Copy the same token into the project server configuration before leaving this page.

Registered credentials: 0

Credential	Status	Expires	Last used	Updated
-	-	-	-	-

From

Task

Entity

Field

Source

Target

Model

Tier

Pricing

The default is the latest 30 calendar days in the project timezone. Current timezone: Europe/Riga. OpenAI actual totals use the imported provider buckets that fall inside the selected period. Press Refresh after changing the period or project.

Press Refresh to load cost data for the selected project.

0 events are missing provider usage or an effective historical tariff, so the total is incomplete.

The calculated project total uses the mapped OpenAI cost export and follows only the selected dates. Task, entity, language, model and pricing filters apply to the table and attributed KPIs, which cover only events recorded with that metadata.

The imported OpenAI cost buckets do not cover the entire selected period. Available coverage: - – -.

No mapped OpenAI cost buckets are available for the selected period. The calculated total is an attributed estimate, not the complete provider spend.

The EUR total is unavailable because the stored ECB EUR/USD history does not cover 0 USD in this result. USD totals are unchanged.

Calculated cost, USD

Calculated cost, EUR

—

ECB daily rates: - – -

OpenAI requests

Input tokens

Output tokens

Attributed cost, USD

0% of project cost

Attributed cost, EUR

—

ECB daily rates

Attributed events

Attributed input

Attributed output

Date	Task / entity	Languages	Model / tier	Mode	Events	Tokens	USD ×1.15	EUR ×1.15	ECB rate	Pricing
-	- -	-	- -	-	0	0	0	—	—	-

Configured tariff versions (0)

Model	Tier	Effective	Max input	Input	Cached	Cache write	Output
-	-	- –	-	0	0	0	0

OpenAI provider projects and Hub mapping

Provider project names come from the official OpenAI Projects directory. Mapping is explicit and does not change immutable Hub events.

OpenAI may return organization-level cost rows without a usable project allocation. Missing provider cost data is not treated as zero; calculated Hub cost continues to come from project events and historical tariffs.

Provider project mapping saved.

Projects: 0; mapped: 0; organization provider cost: 0 USD.

OpenAI project	Status	Requests	Tokens	Provider cost, USD	Hub project
- -	-	0	0	0 —	- Unmapped Suggested:

Storefront Language Jobs

Source is . Existing translations are kept. Pausing jobs does not change CS-Cart language visibility.

Jobs on:

Store language list unavailable:

Loading storefront language jobs...

Storefront languages failed to load:

Storefront	Language	Name	Store availability	Translation jobs
#{{storefront_id}} {{storefront_name}}	{{lang_code}} default	{{name}}	active hidden not active global

Banner media copied to : rows created , links replaced , links inserted .

Banner media sync failed:

Demo banner cleanup for : rows changed , text fields cleared , structural fields copied . Sync warning:

Demo banner cleanup failed:

No storefront target languages loaded.

Storefront language save failed:

Projects / Stores

Project metadata is stored in Postgres. Store DB credentials stay in Wappler connection files.

Project	Store connection
#{{project_id}} {{project_code}} {{name}}	{{store_connection_name}}

Connection OK: , DB , companies .

Connection check failed:

Connection failed:

Project code

Name

Platform

Source

Default target

Store connection

Must match an existing Wappler connection file before the project can be enabled.

Translation connection

Status

SEO automation

SEO products

SEO categories

Media alt automation

Media alt products

Shop URL

Public base URL

Optional public subdomain for callbacks or downloadable files.

Shop API env name

Attachment token env name

Notes

Saved project #.

Cloned settings: entities , product features .

Save failed:

Clone failed:

Project AI Settings

API key is stored encrypted and is never returned by the API. Batch jobs keep the model fixed at creation time.

Key: configured not configured

Default translation model

SEO model

Replace OpenAI API key

Key action

Translation prompt context

SEO project direction/context

Sent to SEO generation and AI image-alt generation.

Prices shown as Batch input/output USD per 1M tokens. Standard API is higher.

AI settings saved.

AI settings save failed:

Storefront Language Jobs

Source is . Existing translations are kept. Pausing jobs does not change CS-Cart language visibility.

Jobs on:

Store language list unavailable:

Storefront	Language	Name	Store availability	Translation jobs
#{{storefront_id}} {{storefront_name}}	{{lang_code}} default	{{name}}	active hidden not active global

Banner media copied to : rows created , links replaced , links inserted .

Banner media sync failed:

Demo banner cleanup for : rows changed , text fields cleared , structural fields copied . Sync warning:

Demo banner cleanup failed:

No storefront target languages loaded.

Storefront language save failed:

Sync From Store

Every apply runs a server-side preflight. Hard unit and queue limits stop the whole extraction before any tr_units upsert.

Entity scope

I reviewed the preflight and explicitly confirm a large extraction

Status . Scope : source rows , normalized units , new units , projected missing/review , missing , stale/check , review , source missing handled (deleted , protected ).

Entity	Status	New	Projected queue	New limit	Queue limit	Confirmation
{{entity_type}}	{{status}}	{{new_units}}	{{projected_pending_review_units}}	{{preflight.max_new_units_per_run}}	{{preflight.max_pending_review_units}}	required not required

Extraction stopped or failed for entities. No units are created when the run status is blocked or confirmation_required.

Translation Settings

Disabled entities are skipped by sync, batch build, and store writes. Product features are opt-in for labels, variants, and custom values.

Entities: . Feature settings enabled: .

Loading translation settings...

Settings load failed:

Entities

No entity settings loaded.

Project	Entity	Model	Global	Project override	Effective	Guardrails	Notes	Updated
{{project_code}} #{{project_id}}	{{entity_type}}		enableddisabled pilot: customer:	enableddisabled pilot_enabled: customer_visible:	effective enabled effective disabled customer:	new/run: {{max_new_units_per_run}} missing/review: {{max_pending_review_units}} warn: {{queue_warning_threshold}}	{{notes}}	{{updated_at}}

Edit project entity override

Project / entity_type

enabled

pilot_enabled

customer_visible

notes

Product Features

Enabled:

Feature	Type	Labels	Variants	Values
#{{feature_id}} {{description}}	{{feature_type}}		-	-

Product features settings are temporarily unavailable:

If no feature is enabled for a column, sync will not create units for that product feature entity.

SEO Automation

Separate SEO flow for active products, categories, static pages, and dynamic pages. It uses only page_title, meta_description, and meta_keywords.

Loading SEO automation...

SEO entity

Limit

Generate source SEO first, write it safely, then translate SEO to the selected target language.
Source language: . Target language: . Products missing SEO: . Categories missing SEO: . Static pages: . Dynamic pages: .

Handled by a separate project cycle: . Other SEO scopes remain active here.

SEO batch # created with items.

Target SEO batch # created with items.

SEO error:

Batch	Task	Entity	Status	Model	Items
#{{seo_batch_id}}	Generate Translate to	{{entity_type}}	ready to write store checked	{{model_name}}	ready to write , written , skipped , failed

Existing translated SEO text is protected. Empty or source-copy SEO text can be replaced. SEO URL/slug fields are not read or written.

Submit all: submitted / new SEO batches. Failed .

Fetch all: processed / unfinished SEO batches. Items: completed , failed . Failed batches .

SEO safe write completed for batch #: written items , written fields , skipped .

SEO batch action failed:

Media Alt Automation

All active storefront images are covered, including linked media and images embedded in localized HTML. Existing translated alt text is protected.

Loading media alt automation...

Limit

Source method

Generate source image alt first, write it safely, then translate alt text to the selected target language.
Source language: . Target language: .

Storefront images	total
Source alt present	(%)
Source alt missing
Ready to batch
In active batch
Selected target needs translation
Decorative, kept empty

Product media is handled by the project's separate product cycle. All other storefront image scopes remain enabled here.

Coverage details

Category images missing source alt: . Blog images missing source alt: . Content HTML image problems: in documents.

: missing / linked images.

HTML: missing / images.

Media alt batch # created with items. Submit and fetch it before safe write.

Target media alt batch # created with items.

Batch	Task	Status	Model	Method	Items
#{{media_alt_batch_id}}	Generate Translate to	ready to write store checked	{{model_name}}	{{generation_method \|\| ''}}	ready to write , written , skipped , failed

Target alt text is trusted only when it matches a result previously confirmed by Translation Hub for the same source alt. Other target values are translated and replaced. This block does not use the standard translation units.

Submit all: submitted / new media alt batches. Failed .

Fetch all: processed / unfinished media alt batches. Items: completed , failed . Failed batches .

Media alt safe write completed for batch #: written items , skipped .

Media alt action failed:

Build AI Batch

Entity

Max items

Max batches

Created batch with items. Mode: . Filled from memory: .

Created check batch with items.

Auto build: created batches / items, memory filled , entities scanned , stop: .

Batch build failed:

Batch Actions

Use row Submit / Fetch, or bulk actions in Batch Monitoring. Re-submit is blocked once a batch is already sent.

Submitted as . Status: .

Batch status: . Fetched: . Items: /, failed , unresolved . OpenAI: /, failed .

Fetch all: processed / unfinished batches. Items: completed , failed , unresolved . OpenAI: completed / , failed .

Submit all: submitted / new batches. Failed .

Batch action failed:

Deleted batch . Reset unresolved units: .

Reset remaining items for batch . Units reset: .

Force retranslate: reset / units, skipped active batches , skipped writes . Next build bypasses memory for these units.

Force products from EN

Resets product title/description fields and product SEO from EN, then creates pending batches. It does not submit batches to OpenAI.

Recovery

Product IDs

Target languages

Standard batch items

SEO batch items

All target languages

Product fields

Only source-changed product fields

SEO fields

Sync store first

Build product batches

Build SEO batches

Explicitly confirm a large product extraction after reviewing counts

Running . Sync, reset, and batch build can take several minutes. Batch table updates after completion.

Mode: . Product fields: actual products found , product-language rows , field units found , fingerprint stale products , fingerprint stale units , selected products , selected product-language rows , selected units , reset products , reset units , batches , items . SEO: product-language rows , existing target SEO , batches , items . Submit pending standard and SEO batches after checking counts.

Job: . Status: . Entity: . Cycles: . Entity cycles: . Remaining entities: . Rounds: . Prepared: . Selected: . Written: . Failed: . Cloned rows: . Stop: . Failure streak: /5. Writing continues in background. Status polling is retrying. Status check is retrying while the backend job continues. Stopped after repeated write failures. Error: .

Mode: . Eligible: . New queue candidates: . Prepared: . Already queued/written: . Use Preview pending, then Write chunk.

Mode: . Pending: . Validated: . Written: . Failed: . Cloned rows: . Limit used: /.

Deleted source-missing units: . Writes: , batch items: , results: .

Failed writes

These rows are in tr_writes with write_status = failed. Retryable rows go back to pending; obsolete rows are skipped so review can create a fresh queue later.

Returned to pending: . Skipped obsolete: .

# / unit /

Loading translation memory metrics...

Memory Rows

Total:

Used From Memory

Rows used:

Manual Overrides

AI accepted:

Last Used

Updated:

Loading coverage summary...

Coverage

Ready:

Units

Translated/Written:

Review Queue

Suspected EN copy:

Batch Progress

Completed: / Failed:

Missing

Batched

Translated

Written

Wrong Language

HTML Problems

Dangerous translation queues

Grouped by project_code, entity_type, target_lang, and status. “Without links” means the unit has no tr_results, tr_writes, or tr_batch_items evidence.

Warning threshold exceeded. Stop extraction and inspect the affected project/entity/language before the queue consumes disk.

Queue monitor failed:

Checking queue sizes and orphan evidence...

Project	Entity	Lang	Status	Rows	Without results	Without writes	Without batch items	Without links	Threshold	Effective
{{project_code}}	{{entity_type}}	{{target_lang}}	{{status}}	{{row_count}}	{{rows_without_results}}	{{rows_without_writes}}	{{rows_without_batch_items}}	{{rows_without_links}}	{{warning_threshold}}	enabled disabled

Coverage By Entity

Loading entity coverage...

Entity coverage failed to load:

Entity	Total	Missing	Review	In batch	Translated	To write	OK in store	Coverage
{{entity_type}}	{{total_units}}	{{missing_units}}	{{review_required_units}}	{{batched_units}}	{{translated_units}}	{{write_candidate_units}}	{{written_units}}	{{coverage_pct}}%

Coverage By Field

Loading field coverage...

Field coverage failed to load:

Entity	Field	Total	Suspected EN	Coverage
{{entity_type}}	{{field_name}}	{{total_units}}	{{suspected_source_copy_units}}	{{coverage_pct}}%

Visible standard batches

Scope: / /

Visible Pending Submitted Completed Failed Other

Submit all: submitted / new batches. Failed .

Submit all failed:

Loading standard batches...

Batch	Entity	Status	Items	External ref	Requested	Sent	Completed
{{batch_id}}	{{entity_type}} entity disabled	{{openai_status \|\| status}} local {{status}}	{{items_completed}} / {{items_total}} pending / sent / failed history / resettable	{{external_batch_ref}}

Review Queue

Unit	Entity	Field	Status	Flags	Target
{{unit_id}}	{{entity_type}} {{entity_key_text}}	{{field_name}}	{{status}}	{{quality_flags_json}}	{{target_value_current}}

Recent Activity

Loading recent activity...

Type	Status	Ref	At
{{activity_type}}	{{status}}	{{ref_id}} {{title}}

Translation Review

Left is EN source. Right is the current target value that will be saved after review.

Inspect any EN to target pair and confirm whether the current target is really translated.

Selected 0 Bulk: selected , approved , reset . Bulk failed:

Selected unit

Status

existing target

Current target equals EN source or is flagged as same_as_source. Use Accept invariant to keep it unchanged, edit the target, or return it to untranslated.

Current target comes from the store source and is not confirmed by this translation pipeline yet. Review or replace it before approval.

Saved unit as .

Reset unit to .

Updated EN source for unit . Affected units: .