@giveitsmaller/contracts 0.82.0 → 0.84.0

package/README.md

@@ -5,8 +5,8 @@
 
 | | |
 |---|---|
-| Package version | `0.82.0` |
-| Generated from spec | **`v2.217.0`** |
+| Package version | `0.84.0` |
+| Generated from spec | **`v2.219.0`** |
 
 ## Two version lines, and they are not comparable
 

package/accepted-options/accepted-options.json

@@ -1214,5 +1214,5 @@
       }
     }
   },
-  "schema_version": "2.217.0"
+  "schema_version": "2.219.0"
 }

package/accepted-options/image-output-routes.json

@@ -260,5 +260,5 @@
       "source_op": "compress"
     }
   },
-  "schema_version": "2.217.0"
+  "schema_version": "2.219.0"
 }

package/anonymous-policy.yaml

@@ -1,6 +1,14 @@
 # Anonymous (account-less) policy — the ONE declaration of what a guest may do
-# (ticket UQ7zROTr; sdks u6Q9oxuI follow-up; owner decision 610(4), 2026-09-26:
-# guests stay images-only with the daily limit).
+# (ticket UQ7zROTr; sdks u6Q9oxuI follow-up).
+#
+# 2.1.0 (card gpYKURQd, hub decision 638(a)): `video.probe_pending_recovery`
+# and `video.probe_pending_counts_against_limits`, additive.
+#
+# 2.0.0 (card 43dDKeEh; owner decision + hub ruling 2026-09-26 21:26): guests
+# move from "30 image jobs per day" to a CREDIT allowance, for all file types,
+# with video capped at 60 s. Shape agreed with api (msg cc5bedfe). BREAKING for
+# readers of 1.0.0: `upload.mime_prefixes` and `quota.workflow_creates_per_ip_per_24h`
+# are gone; read `upload.media` and `allowance` instead.
 #
 # Read by: the SDKs (gisl.anonymous() / Gisl::anonymous() pin their surface to
 # it), api (pins its constants to it: AnonymousOperationPolicy,
@@ -8,34 +16,63 @@
 # tests/test_anonymous_policy.py (holds the relations below).
 #
 # The refusals are declared in openapi `POST /api/workflows` 403
-# (AnonymousOperationNotAllowedResponse / AnonymousQuotaExhaustedResponse);
-# this file declares the POLICY that produces them, so a client can predict
-# the 403 instead of learning it.
+# (AnonymousOperationNotAllowedResponse / AnonymousQuotaExhaustedResponse /
+# AnonymousLimitExceededResponse); this file declares the
+# POLICY that produces them, so a client can predict the refusal instead of
+# learning it.
 
-schema_version: "1.0.0"
+schema_version: "2.1.0"
 
 # Operations an anonymous caller may request at workflow-create. Anything else
 # answers 403 `anonymous_operation_not_allowed`.
 operations_allowed: [compress, convert, thumbnail]
 
 upload:
-  # Only images: any `image/*` MIME.
-  mime_prefixes: ["image/"]
+  # ALL file types: every MIME the upload endpoints accept (was images only).
+  media: all
   # THE guest per-file cap: 10,000,000 B, the single-shot endpoint's own maximum
   # (openapi UploadThresholds.single_shot_max_bytes). Guests upload single-shot
-  # only, so a larger cap would be unreachable. api lowers its
-  # AnonymousUploadPolicy from 10 MiB to this and pins to this file (xOaNkDll,
-  # 2026-09-26; owner decision 610(4): "images only, 10 MB").
+  # only, so a larger cap would be unreachable (xOaNkDll).
   max_file_bytes: 10000000
   # Multipart initiate/status/presign/keepalive require authentication (api
   # security.yaml + a defensive 401 in initiate); complete is unreachable
   # without an initiated session.
   multipart_allowed: false
 
-quota:
-  # Rolling 24 h, per client IP. The 31st create answers 403 `anonymous_quota_exhausted`.
-  workflow_creates_per_ip_per_24h: 30
-  # Per-minute rate limits (429 when exceeded).
-  per_minute:
-    upload: 5
-    workflow_create: 2
+video:
+  # Measured on the UPLOAD PROBE's duration. A guest video create waits for the
+  # probe (the existing probe-pending gate), then is refused above this with
+  # 403 `anonymous_limit_exceeded` (limit `video_duration_seconds`).
+  max_duration_seconds: 60
+  exceeded_error_type: anonymous_limit_exceeded
+  # Until the probe lands, a guest video create is refused 422 `probe_pending`
+  # (hub decision 638(a), contract-first for api #795). A guest RETRIES THE
+  # CREATE, on `Retry-After` when sent, else its own backoff: the probe poll
+  # endpoint (POST /api/uploads/{id}/probe) requires sign-in, so a guest
+  # cannot poll it. That refusal does not count against `per_minute.
+  # workflow_create` and reserves no `allowance` credits.
+  probe_pending_recovery: retry_create
+  probe_pending_counts_against_limits: false
+
+allowance:
+  # Credits per client IP over a ROLLING 24 h (a sliding window, so there is no
+  # calendar-day or timezone edge). Identity is the IP alone: a funnel nudge,
+  # not a security boundary (NAT shares it).
+  credits: 30
+  window: rolling_24h
+  identity: ip
+  # Priced with the same per-operation weights as a signed-in caller
+  # (`operation_weights.yaml`), charged at workflow create against the IP's
+  # rolling ledger, and released on failure as for a signed-in reservation.
+  priced_by: operation_weights
+  # A create whose cost would cross the allowance is refused WHOLE (no partial
+  # run) with 403 `anonymous_quota_exhausted`, carrying `allowance`,
+  # `remaining` and `cost`. `Retry-After` (seconds until enough of the window
+  # frees up) is sent only when `cost <= credits`; a request costing more than
+  # the whole allowance can never fit and gets none.
+  exhausted_error_type: anonymous_quota_exhausted
+
+# Per-minute rate limits (429 when exceeded), unchanged from 1.0.0.
+per_minute:
+  upload: 5
+  workflow_create: 2

package/asyncapi/README.md

@@ -9,7 +9,7 @@
 - **Spec:** `compression_contracts/asyncapi/events.yaml` (AsyncAPI 3.0)
 - **Generator:** [asyncapi/cli](https://www.asyncapi.com/tools/cli) 6.0.0
   (Docker image: `asyncapi/cli:6.0.0`)
-- **Files:** 30
+- **Files:** 31
 
 ## Contents
 

package/asyncapi/events.yaml

@@ -1570,10 +1570,13 @@ channels:
       here directly (NOT via SNS); a terraform-owned launcher reads the spec from
       S3 (`pointer`) and runs a Fargate task. `max_receive_count = 2` (ADR-0007).
 
-      On this queue the publisher always sets `operation_type: merge` (shared
-      `LongFormJobMessage` enum permits `{compress, merge}`; per-queue
-      `operation_type` is a publisher invariant documented here, not a separate
-      schema).
+      On this queue the publisher always sets `operation_type: merge` (the
+      shared `LongFormJobMessage` enum permits `{compress, merge, split}`;
+      per-queue `operation_type` is a publisher invariant documented here, not a
+      separate schema). `processing_class` is `long_form` for a re-encode merge
+      and `long_form_concat` for a stream-copy join (a fan-out rejoin, or a
+      user merge with `re_encode_mode: never`); the launcher routes the second
+      to its stream-copy task.
     bindings:
       sqs:
         queue:
@@ -2459,8 +2462,10 @@ operations:
         bindingVersion: '0.2.0'
     summary: Publish a long-form merge re-encode claim-check.
     description: |
-      As `publishLongFormCompressVideo`, for `operation_type = merge`
-      (`merge.video.long_form_re_encode`). Direct-to-SQS claim-check; no SNS.
+      As `publishLongFormCompressVideo`, for `operation_type = merge`.
+      `processing_class` selects the task: `long_form` for a re-encode merge
+      (`merge.video.long_form_re_encode`), `long_form_concat` for a stream-copy
+      join. Direct-to-SQS claim-check; no SNS.
     messages:
       - $ref: '#/channels/longformMergeVideoReencode/messages/longFormJob'
 
@@ -2517,9 +2522,10 @@ operations:
         bindingVersion: '0.2.0'
     summary: Consume a long-form merge re-encode claim-check.
     description: |
-      As `consumeLongFormCompressVideo`, for `operation_type = merge`
-      (`merge.video.long_form_re_encode`). The launcher resolves the spec
-      from S3 via `pointer` and runs the merge-re-encode Fargate task.
+      As `consumeLongFormCompressVideo`, for `operation_type = merge`. The
+      launcher resolves the spec from S3 via `pointer` and forks on
+      `processing_class`: `long_form` runs the merge-re-encode Fargate task,
+      `long_form_concat` the small stream-copy concat task.
     messages:
       - $ref: '#/channels/longformMergeVideoReencode/messages/longFormJob'
 
@@ -2920,6 +2926,14 @@ components:
             operation_id: "018f9c42-5d6b-7481-b3df-9fd0a0a5e201"
             operation_type: "merge"
             processing_class: "long_form"
+        - name: Long-form stream-copy join
+          summary: merge.video.long_form_concat claim-check (fan-out rejoin, or re_encode_mode never)
+          payload:
+            pointer: "job-specs/018f9c42-5d6b-7481-b3df-9fd0a0a5e114.json"
+            job_id: "018f9c42-5d6b-7481-b3df-9fd0a0a5e114"
+            operation_id: "018f9c42-5d6b-7481-b3df-9fd0a0a5e202"
+            operation_type: "merge"
+            processing_class: "long_form_concat"
 
     # ============================================
     # OPERATION PROGRESS MESSAGE
@@ -3915,6 +3929,23 @@ components:
         consumers validating against this schema MUST tolerate additional or
         differently-shaped redriven messages. The producer contract this schema
         pins is the **required** set.
+
+        **`processing_class: long_form_concat` requires `operation_type:
+        merge`** (the `if`/`then` below): the launcher diverts only that pair.
+      # Deployed launcher verified 2026-09-26 (code read back from AWS, staging
+      # and prod): handler.py `elif operation_type == "merge" and
+      # processing_class == "long_form_concat"` -> concat_streamcopy taskdef
+      # (2 vCPU / 4 GB, ACTIVE in both accounts).
+      if:
+        properties:
+          processing_class:
+            const: long_form_concat
+        required:
+          - processing_class
+      then:
+        properties:
+          operation_type:
+            const: merge
       required:
         - pointer
         - job_id
@@ -4795,14 +4826,23 @@ components:
         Operation-specific metrics. Content varies by operation type.
         All metrics include duration_ms. Compression includes size/ratio.
         Merge includes input count. Archive includes file count and total size.
-      # estimated_saving_pct only means something on an already-optimal result.
-      if:
-        required: [estimated_saving_pct]
-      then:
-        required: [already_optimal]
-        properties:
-          already_optimal:
-            const: true
+      # estimated_saving_pct and already_optimal_kind only mean something on an
+      # already-optimal result.
+      allOf:
+        - if:
+            required: [estimated_saving_pct]
+          then:
+            required: [already_optimal]
+            properties:
+              already_optimal:
+                const: true
+        - if:
+            required: [already_optimal_kind]
+          then:
+            required: [already_optimal]
+            properties:
+              already_optimal:
+                const: true
       properties:
         original_size_bytes:
           type: integer
@@ -4969,8 +5009,43 @@ components:
             because compressing it would not have made it smaller. Set by the
             compression processor itself, never by an individual worker's
             extras. Absent or `false` on an ordinary compression; consumers
-            treat the two the same.
+            treat the two the same. Required of EVERY same-format compress
+            path (resolved `same_format` route: `output_format` original,
+            absent, or equal to the input's format), short- and long-form (a long-form worker copies the source
+            to the output key; ⚠️ not yet shipped, so today only short-form
+            conforms), instead of `output_too_large`. TARGET billing: the API
+            charges nothing for the operation (owner decision 645(4),
+            `schemas/pricing-policy.yaml`); TODAY it still charges it
+            (`charge.not_charged_api_today`) until api ships.
+            **Location: THIS declared field, `metrics.already_optimal`.**
+            Transition (expand/migrate, lambdas `KiJCEsW2`): short-form
+            workers have emitted it only inside the open-map extras, which the
+            deployed worker SDK serialises as a NESTED object literally keyed
+            `additionalProperties` (camelCase) — a distinct JSON path,
+            `metrics.additionalProperties.already_optimal`, never the same key
+            as this field (api `OperationResultMetadataProjection` reads
+            flat first, then that nested object). When both are present and
+            disagree, THIS field wins. Workers set BOTH until every path sets this one; the API
+            reads this field OR the extras key, and its billing reads the same
+            place its `/status` projection reads (once billing reads it at all),
+            so the refund and the badge cannot disagree. The extras key is then dropped.
           example: true
+        already_optimal_kind:
+          type: string
+          enum: [not_smaller]
+          description: |
+            OPTIONAL, only with `already_optimal: true`: why the original was
+            returned. `not_smaller` = compressed, and the result was not
+            smaller: there is no quality trade-off to explain. It is the only
+            kind any worker produces (no production path declines to encode an
+            efficient source, lambdas 2026-09-30), so no other value is
+            declared; a consumer shows neutral "already optimised" copy for an
+            absent kind and for any value it does not know. The API
+            projects it onto the REST `OperationResultMetadata` (same name and
+            values). Contract-first (card `cDZuuyg8`). Deliberately NOT named
+            `already_optimal_reason`: that name is a worker diagnostic the API
+            must not serve (roNRMilt); this is a closed, user-facing kind.
+          example: not_smaller
         estimated_saving_pct:
           type: number
           minimum: 0
@@ -5818,21 +5893,36 @@ components:
       type: string
       description: |
         Logical processing class on the long-form claim-check wire
-        (`LongFormJobMessage.processing_class`) — always the literal
-        `long_form` (never `short_form` / `blocked`; those exist only on the
-        upload-probe response and HTTP plan surfaces). The launcher forks on
-        this field; do not confuse it with the backend worker identity, which
-        stays off the wire per `schemas/FORMAT.md` §F8.
-
-        A single-value enum (not a JSON-Schema `const`): named + `enum` is the
-        repo's proven typed-codegen form (mirrors `JobInputRole` /
-        `ReEncodeDecision`), whereas `const` degraded to an unconstrained
-        `String` in SDK generators. Lifted from inline at
+        (`LongFormJobMessage.processing_class`). Never `short_form` /
+        `blocked`: those exist only on the upload-probe response and HTTP plan
+        surfaces. The launcher forks on this field together with
+        `operation_type`; do not confuse it with the backend worker identity,
+        which stays off the wire per `schemas/FORMAT.md` §F8.
+
+        - `long_form`: every long-form job except the one below.
+        - `long_form_concat` (card
+          [`JzS0lkSr`](https://trello.com/c/JzS0lkSr), api `m9Bstzaa`): a
+          STREAM-COPY join, valid ONLY with `operation_type: merge` (enforced on
+          `LongFormJobMessage`). The API sends it for (a) the stream-copy rejoin
+          of a long-form fan-out and (b) a user long-form merge with
+          `re_encode_mode: never`. The launcher routes the pair
+          `(merge, long_form_concat)` to its small stream-copy task instead of
+          the merge re-encode task. Any other pair would fall through to that
+          operation's default task with a class it does not know, which is why
+          the pairing is a schema rule and not a convention.
+
+        Named + `enum`, not `const`: the repo's proven typed-codegen form
+        (mirrors `JobInputRole` / `ReEncodeDecision`); `const` degraded to an
+        unconstrained `String` in SDK generators. Lifted from inline at
         `LongFormJobMessage.processing_class` per ticket
-        [`A3DyuYUd`](https://trello.com/c/A3DyuYUd). Single-member enum
-        validates identically to the prior `const long_form`.
+        [`A3DyuYUd`](https://trello.com/c/A3DyuYUd).
+      # CLOSED (codex #625): a routing discriminator. An unknown class would
+      # select an unintended launcher path, so a new value reaches consumers
+      # deliberately, never by tolerance.
+      x-enum-openness: closed
       enum:
         - long_form
+        - long_form_concat
 
     ProgressStatus:
       type: string
@@ -5970,7 +6060,7 @@ components:
         - format_mismatch: MIME type doesn't match content
         - decode_failed: Cannot decode/parse file
         - processing_failed: Processing library error
-        - output_too_large: Output exceeded the size limit
+        - output_too_large: Output exceeded the size limit. NOT emitted by a compress on the resolved same_format route (output_format original, absent, or equal to the input's format, e.g. webp to webp), which returns the original with OperationMetrics.already_optimal instead (owner decision 645(4)). Today short-form only: long-form compress still emits it until lambdas ships its copy path
         - invalid_options: Invalid operation options
         - missing_source: Source file not found in storage
         - invalid_key: Storage key is malformed

package/availability/availability.json

@@ -1,6 +1,6 @@
 {
   "capabilities_profile": "baseline",
-  "capabilities_version": 231,
+  "capabilities_version": 233,
   "endpoints": {
     "DELETE /api/auth/account": {
       "auth": "required",
@@ -2060,7 +2060,7 @@
             },
             "audio_codec": {
               "default": "copy",
-              "description": "Audio codec. copy = passthrough (keep the source stream as-is; re-encode only on explicit request) \u2014 aac/opus/vorbis re-encode. Default is `copy`: on the same-container path (`output_format` original/absent) the source stream is already container-compatible, so copy is safe. When `output_format` RE-TARGETS the container, an absent/default audio_codec resolves to a container-valid codec via the facade (e.g. webm -> opus), NOT a blind copy of an incompatible stream \u2014 a DEFAULT is never rejected (the exact per-container resolution, incl. ogg vorbis-vs-opus, is an output_format stable-flip prerequisite). A caller-EXPLICIT codec the resolved container cannot carry (e.g. aac on a webm target) is rejected.",
+              "description": "Audio codec. copy = passthrough (keep the source stream as-is; re-encode only on explicit request) \u2014 aac/opus/vorbis re-encode. Default is `copy`: on the same-container path (`output_format` original/absent) the source stream is already container-compatible, so copy is safe. When `output_format` RE-TARGETS the container, an absent/default audio_codec resolves to a container-valid codec via the facade (e.g. webm -> opus), NOT a blind copy of an incompatible stream \u2014 a DEFAULT is never rejected (the exact per-container resolution, incl. ogg vorbis-vs-opus, is an output_format stable-flip prerequisite). `copy` sent explicitly counts as the default: if the source audio cannot be copied into the target container it is re-encoded to the container default (webm -> opus) and reported in dropped_options. A NON-default codec the resolved container cannot carry (e.g. aac on a webm target) is rejected.",
               "type": "enum",
               "values": [
                 "aac",
@@ -2071,7 +2071,7 @@
             },
             "codec": {
               "default": "h264",
-              "description": "Video codec. h264 = widest compatibility, h265/av1 = better compression but slower. The default (h264) is the MP4/MOV standard and is CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input container): for a container that cannot carry it (e.g. WebM, which carries VP8/VP9/AV1) the server resolves a container-valid codec (WebM -> vp9). A DEFAULT value is never rejected \u2014 including an explicit `output_format: webm` left at the default `h264` (resolves to vp9, not rejected); only a caller-EXPLICIT incompatible codec is rejected.",
+              "description": "Video codec. h264 = widest compatibility, h265/av1 = better compression but slower. The default (h264) is the MP4/MOV standard and is CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input container): for a container that cannot carry it (e.g. WebM, which carries VP8/VP9/AV1) the server resolves a container-valid codec (WebM -> vp9). A DEFAULT value is never rejected \u2014 including an explicit `output_format: webm` left at the default `h264` (resolves to vp9, not rejected). A value equal to the default counts as the default even when sent, and the substitution is reported in dropped_options. Only a NON-default codec the container cannot carry is rejected (e.g. h265 on WebM).",
               "type": "enum",
               "values": [
                 "h264",
@@ -2101,7 +2101,7 @@
             },
             "faststart": {
               "default": true,
-              "description": "Move the MP4/MOV moov atom to the front so the file starts playing before it is fully downloaded (progressive web playback). MP4/MOV only. CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input): the default (true) is silently inapplicable for a container with no moov atom (e.g. WebM) \u2014 that is NOT a rejection. Only a caller-EXPLICIT faststart:true on such a container is rejected.",
+              "description": "Move the MP4/MOV moov atom to the front so the file starts playing before it is fully downloaded (progressive web playback). MP4/MOV only. CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input): on a container with no moov atom (e.g. WebM) the default (true), omitted or sent, is dropped and reported in dropped_options \u2014 never a rejection. false there asks for nothing and is accepted.",
               "type": "boolean"
             },
             "fit": {
@@ -2134,7 +2134,7 @@
             },
             "output_format": {
               "default": "original",
-              "description": "Output container. original = keep the input container (default, same-format compression). Other values change the container as part of the same job, keeping the size cap. When set, the RESOLVED target container determines the valid codec/audio_codec/faststart sets \u2014 an absent/default codec resolves to a container-valid one (e.g. webm -> vp9) and is never rejected; only a caller-EXPLICIT incompatible value is rejected.",
+              "description": "Output container. original = keep the input container (default, same-format compression). Other values change the container as part of the same job, keeping the size cap. When set, the RESOLVED target container determines the valid codec/audio_codec/faststart sets \u2014 an absent or default-equal value (sent equal to its default counts as the default) resolves to a container-valid one (e.g. webm -> vp9), is never rejected, and is reported in dropped_options; only a NON-default incompatible value is rejected.",
               "per_value_availability": {
                 "mp4": {
                   "availability": "planned"
@@ -5498,7 +5498,7 @@
       "sole_op": true
     }
   },
-  "schema_version": "2.217.0",
+  "schema_version": "2.219.0",
   "source_commit": null,
   "user_tier": null,
   "workflow_features": {

package/code-builder/code-builder-metadata.json

@@ -2824,7 +2824,7 @@
             },
             "audio_codec": {
               "default": "copy",
-              "description": "Audio codec. copy = passthrough (keep the source stream as-is; re-encode only on explicit request) \u2014 aac/opus/vorbis re-encode. Default is `copy`: on the same-container path (`output_format` original/absent) the source stream is already container-compatible, so copy is safe. When `output_format` RE-TARGETS the container, an absent/default audio_codec resolves to a container-valid codec via the facade (e.g. webm -> opus), NOT a blind copy of an incompatible stream \u2014 a DEFAULT is never rejected (the exact per-container resolution, incl. ogg vorbis-vs-opus, is an output_format stable-flip prerequisite). A caller-EXPLICIT codec the resolved container cannot carry (e.g. aac on a webm target) is rejected.",
+              "description": "Audio codec. copy = passthrough (keep the source stream as-is; re-encode only on explicit request) \u2014 aac/opus/vorbis re-encode. Default is `copy`: on the same-container path (`output_format` original/absent) the source stream is already container-compatible, so copy is safe. When `output_format` RE-TARGETS the container, an absent/default audio_codec resolves to a container-valid codec via the facade (e.g. webm -> opus), NOT a blind copy of an incompatible stream \u2014 a DEFAULT is never rejected (the exact per-container resolution, incl. ogg vorbis-vs-opus, is an output_format stable-flip prerequisite). `copy` sent explicitly counts as the default: if the source audio cannot be copied into the target container it is re-encoded to the container default (webm -> opus) and reported in dropped_options. A NON-default codec the resolved container cannot carry (e.g. aac on a webm target) is rejected.",
               "enum_ref": "AudioCodec",
               "enum_translations": {
                 "aac": "Aac",
@@ -2851,7 +2851,7 @@
             },
             "codec": {
               "default": "h264",
-              "description": "Video codec. h264 = widest compatibility, h265/av1 = better compression but slower. The default (h264) is the MP4/MOV standard and is CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input container): for a container that cannot carry it (e.g. WebM, which carries VP8/VP9/AV1) the server resolves a container-valid codec (WebM -> vp9). A DEFAULT value is never rejected \u2014 including an explicit `output_format: webm` left at the default `h264` (resolves to vp9, not rejected); only a caller-EXPLICIT incompatible codec is rejected.",
+              "description": "Video codec. h264 = widest compatibility, h265/av1 = better compression but slower. The default (h264) is the MP4/MOV standard and is CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input container): for a container that cannot carry it (e.g. WebM, which carries VP8/VP9/AV1) the server resolves a container-valid codec (WebM -> vp9). A DEFAULT value is never rejected \u2014 including an explicit `output_format: webm` left at the default `h264` (resolves to vp9, not rejected). A value equal to the default counts as the default even when sent, and the substitution is reported in dropped_options. Only a NON-default codec the container cannot carry is rejected (e.g. h265 on WebM).",
               "enum_ref": "VideoCodec",
               "enum_translations": {
                 "av1": "Av1",
@@ -2907,7 +2907,7 @@
             },
             "faststart": {
               "default": true,
-              "description": "Move the MP4/MOV moov atom to the front so the file starts playing before it is fully downloaded (progressive web playback). MP4/MOV only. CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input): the default (true) is silently inapplicable for a container with no moov atom (e.g. WebM) \u2014 that is NOT a rejection. Only a caller-EXPLICIT faststart:true on such a container is rejected.",
+              "description": "Move the MP4/MOV moov atom to the front so the file starts playing before it is fully downloaded (progressive web playback). MP4/MOV only. CONTAINER-CONDITIONAL against the RESOLVED output container (the explicit `output_format` if set, else the input): on a container with no moov atom (e.g. WebM) the default (true), omitted or sent, is dropped and reported in dropped_options \u2014 never a rejection. false there asks for nothing and is accepted.",
               "origin": "contract",
               "sdk_exposure": "expose",
               "surface_key": "faststart",
@@ -2965,7 +2965,7 @@
             },
             "output_format": {
               "default": "original",
-              "description": "Output container. original = keep the input container (default, same-format compression). Other values change the container as part of the same job, keeping the size cap. When set, the RESOLVED target container determines the valid codec/audio_codec/faststart sets \u2014 an absent/default codec resolves to a container-valid one (e.g. webm -> vp9) and is never rejected; only a caller-EXPLICIT incompatible value is rejected.",
+              "description": "Output container. original = keep the input container (default, same-format compression). Other values change the container as part of the same job, keeping the size cap. When set, the RESOLVED target container determines the valid codec/audio_codec/faststart sets \u2014 an absent or default-equal value (sent equal to its default counts as the default) resolves to a container-valid one (e.g. webm -> vp9), is never rejected, and is reported in dropped_options; only a NON-default incompatible value is rejected.",
               "origin": "contract",
               "per_value_availability": {
                 "mp4": {
@@ -7643,7 +7643,7 @@
     }
   },
   "preset_config_hash": "sha256:354814f7906f85be6a6ca5ede6c0722c83cdc058fd658a9bb2b0996520466c6d",
-  "schema_version": "2.217.0",
-  "sdk_spec_version": "2.10.0",
-  "source_hash": "sha256:ceb884bc449ef43132183525188505bfcb6b5db4dfb49335a5118844444ddcad"
+  "schema_version": "2.219.0",
+  "sdk_spec_version": "2.11.0",
+  "source_hash": "sha256:62bdaf53d3871346d4e2914e9c9035bf7b0b7c6b03c030eaea12b7854244de14"
 }

package/dist/asyncapi/AnonymousSchema_286.d.ts

@@ -0,0 +1,4 @@
+declare enum AnonymousSchema_286 {
+    NOT_SMALLER = "not_smaller"
+}
+export default AnonymousSchema_286;

package/dist/asyncapi/AnonymousSchema_286.js

@@ -0,0 +1,5 @@
+var AnonymousSchema_286;
+(function (AnonymousSchema_286) {
+    AnonymousSchema_286["NOT_SMALLER"] = "not_smaller";
+})(AnonymousSchema_286 || (AnonymousSchema_286 = {}));
+export default AnonymousSchema_286;

package/dist/asyncapi/LongFormProcessingClass.d.ts

@@ -1,4 +1,5 @@
 declare enum LongFormProcessingClass {
-    LONG_FORM = "long_form"
+    LONG_FORM = "long_form",
+    LONG_FORM_CONCAT = "long_form_concat"
 }
 export default LongFormProcessingClass;

package/dist/asyncapi/LongFormProcessingClass.js

@@ -1,5 +1,6 @@
 var LongFormProcessingClass;
 (function (LongFormProcessingClass) {
     LongFormProcessingClass["LONG_FORM"] = "long_form";
+    LongFormProcessingClass["LONG_FORM_CONCAT"] = "long_form_concat";
 })(LongFormProcessingClass || (LongFormProcessingClass = {}));
 export default LongFormProcessingClass;

package/dist/asyncapi/OperationMetrics.d.ts

@@ -1,4 +1,5 @@
 import ReEncodeDecision from './ReEncodeDecision.js';
+import AnonymousSchema_286 from './AnonymousSchema_286.js';
 interface OperationMetrics {
     'original_size_bytes'?: number;
     'output_size_bytes'?: number;
@@ -17,6 +18,7 @@ interface OperationMetrics {
     'memory_mb'?: number;
     'fargate_runtime_ms'?: number;
     'already_optimal'?: 'completed' | boolean;
+    'already_optimal_kind'?: AnonymousSchema_286;
     'estimated_saving_pct'?: number;
     'additionalProperties'?: Record<string, any>;
 }

package/dist/asyncapi/index.d.ts

@@ -1,4 +1,5 @@
 export { default as AnonymousSchema_217 } from './AnonymousSchema_217.js';
+export { default as AnonymousSchema_286 } from './AnonymousSchema_286.js';
 export { default as ErrorCode } from './ErrorCode.js';
 export { default as Failure } from './Failure.js';
 export { default as JobInputRole } from './JobInputRole.js';

package/dist/asyncapi/index.js

@@ -1,4 +1,5 @@
 export { default as AnonymousSchema_217 } from './AnonymousSchema_217.js';
+export { default as AnonymousSchema_286 } from './AnonymousSchema_286.js';
 export { default as ErrorCode } from './ErrorCode.js';
 export { default as JobInputRole } from './JobInputRole.js';
 export { default as LongFormOperationType } from './LongFormOperationType.js';

package/dist/openapi/models/AccountLimitEntry.d.ts

@@ -2,7 +2,7 @@
  * GISL Compression API
  * REST API for the GISL (Give It Smaller) file compression and processing service.  **Architecture:** - Upload files to get a `file_id` - Create workflows referencing uploaded files with operations (compress, thumbnail, image_watermark, text_watermark, merge, archive, convert, custom_luma, audio_overlay, audio_watermark) - Poll status, stream SSE events, or receive webhook callbacks - Download results per operation output  **Response envelope:** All mutation and query endpoints return `{ success: true, data: {...} }` on success and `{ success: false, error: \"...\", details: [...] }` on failure. Exceptions: `GET /api/operations/schema` returns raw JSON (per-tier private caching with ETag revalidation per ADR-0002 + I3), health probes return flat objects, and `POST /api/contact` returns 204 with no body.  **Availability metadata.** This spec uses the `x-availability` vendor extension as **decorative documentation only**. Per [ADR-0001](../docs/decisions/0001-contract-first-availability.md) §1.5, the runtime endpoint `GET /api/operations/schema` (ticket I3) is the authoritative source; the sidecar `availability.json` (ticket I3b) is the authoritative companion (generated, never hand-edited; CI cross-checks runtime ⇄ sidecar). SDKs MUST NOT depend on `x-availability` reaching generated code — code-generators that surface vendor extensions may emit it as documentation, but consumers read availability from the runtime endpoint, not from the generated bindings.  The 5-value vocabulary (`stable | beta | experimental | planned | deprecated`) is defined in the `AvailabilityValue` schema. See `schemas/FORMAT.md` §Availability Taxonomy for the operational rules (parser obligation: absent = stable; per-enum-value granularity is the `per_value_availability` primitive landed via ticket I17).  **Free-text string fields: `x-string-vocabulary` (ticket [`Q79yjcFF`](https://trello.com/c/Q79yjcFF)).** A `type: string` field with no `enum` that names example values says, as data, what a client may do with them (the same marker is used in the AsyncAPI document): - `open` — a vocabulary that grows. Each published value keeps its   meaning, the SET is not closed: switch on the values you know and   handle an unknown one as the generic case (e.g. `ErrorEnvelope.error`). - `advisory` — explanatory text. Display or log it; **never switch on   it** (e.g. `SseWorkflowTerminalData.reason`). - `none` — not a vocabulary at all (an expression or an identifier,   e.g. `OptionSchema.pattern`). A field whose description hedges with \"common values\" or \"free-form\" must carry the marker; a test enforces it.  **Localisation (per ticket [I26](https://trello.com/c/rcnqwgI4)).**  Error responses + paused/blocked workflow statuses carry a localised human-readable `message` alongside a stable, never-localised `message_key`. Machine-readable fields (`error`, enum values, status codes) stay canonical English.  - **Currently committed locales:** `en-GB` only (per ticket   [`4GKyuYo6`](https://trello.com/c/4GKyuYo6)). The I26 carrier   shape (`Accept-Language` + `Content-Language` + `Vary` headers +   `locale` envelope field + `message_key` + `message_params`) is   stable and exercised; the **catalog** of translated `message`   strings is en-GB-only at runtime today. Additional locales (e.g.   `pt-PT`) will be advertised by name when their catalogs ship —   the request/response carrier shape does NOT change when a new   locale lands. Treat unrequested locales as \"machine-code +   `message_key` path is committed; localised `message` prose is   not\" until this prose enumerates them by name. - **Request:** `Accept-Language` header per RFC 9110 §12.5.4 (q-value   negotiation supported). The server selects the best-match locale   from its supported list; falls back to `en-GB` when no match —   which, until additional catalogs land, is every non-`en-GB`   `Accept-Language`. - **Response:** `Content-Language: <locale>` echo on every localised   response; `Vary: Accept-Language` on every response (CDN/cache   correctness — different `Accept-Language` requests produce   different responses). `Vary` is emitted unconditionally so the   header contract does not flip when a second locale ships. - **Fallback locale:** `en-GB` (also the canonical locale for   `message_key` translations and English `message` prose). - **SDK guidance:** switch on `error` (machine code) for typed   error branches; surface `message_key` to client-side i18n   catalogs (SDK companion work tracked at X19, cross-repo);   display `message` for end-user UI; **never parse `message` for   control flow** — it changes per locale.  Carrier shape lives on `ErrorEnvelope` (envelope-level optional `message_key` + `message` + `locale` + `message_params`) and `ValidationErrorEnvelope` (also per-`details[]` entry). Existing 402 / 403 / 422 envelopes (`BalanceExhaustedResponse`, `FeatureNotAvailableResponse`, `FeatureTierRestrictedResponse`, `WorkflowPausedDetail`) inherit the convention.  **Upload thresholds (per tickets [u0ar7Yye](https://trello.com/c/u0ar7Yye) + [58nBQLWQ](https://trello.com/c/58nBQLWQ)).** Canonical upload constants (single-shot cap, multipart chunk size, multipart concurrency default, multipart first-chunk size) live on the `UploadThresholds` schema with `const:`-pinned values. SDK generators emit these as typed binding constants so frontend / API / SDKs reference one source of truth instead of hardcoding magic numbers. A runtime `GET /api/uploads/limits` endpoint for dynamic discovery (per-tier / per-environment overrides) is a deferred follow-up.
  *
- * The version of the OpenAPI document: 2.217.0
+ * The version of the OpenAPI document: 2.219.0
  *
  *
  * NOTE: This class is auto generated by OpenAPI Generator (https://openapi-generator.tech).

package/dist/openapi/models/AccountLimitEntry.js

@@ -4,7 +4,7 @@
  * GISL Compression API
  * REST API for the GISL (Give It Smaller) file compression and processing service.  **Architecture:** - Upload files to get a `file_id` - Create workflows referencing uploaded files with operations (compress, thumbnail, image_watermark, text_watermark, merge, archive, convert, custom_luma, audio_overlay, audio_watermark) - Poll status, stream SSE events, or receive webhook callbacks - Download results per operation output  **Response envelope:** All mutation and query endpoints return `{ success: true, data: {...} }` on success and `{ success: false, error: \"...\", details: [...] }` on failure. Exceptions: `GET /api/operations/schema` returns raw JSON (per-tier private caching with ETag revalidation per ADR-0002 + I3), health probes return flat objects, and `POST /api/contact` returns 204 with no body.  **Availability metadata.** This spec uses the `x-availability` vendor extension as **decorative documentation only**. Per [ADR-0001](../docs/decisions/0001-contract-first-availability.md) §1.5, the runtime endpoint `GET /api/operations/schema` (ticket I3) is the authoritative source; the sidecar `availability.json` (ticket I3b) is the authoritative companion (generated, never hand-edited; CI cross-checks runtime ⇄ sidecar). SDKs MUST NOT depend on `x-availability` reaching generated code — code-generators that surface vendor extensions may emit it as documentation, but consumers read availability from the runtime endpoint, not from the generated bindings.  The 5-value vocabulary (`stable | beta | experimental | planned | deprecated`) is defined in the `AvailabilityValue` schema. See `schemas/FORMAT.md` §Availability Taxonomy for the operational rules (parser obligation: absent = stable; per-enum-value granularity is the `per_value_availability` primitive landed via ticket I17).  **Free-text string fields: `x-string-vocabulary` (ticket [`Q79yjcFF`](https://trello.com/c/Q79yjcFF)).** A `type: string` field with no `enum` that names example values says, as data, what a client may do with them (the same marker is used in the AsyncAPI document): - `open` — a vocabulary that grows. Each published value keeps its   meaning, the SET is not closed: switch on the values you know and   handle an unknown one as the generic case (e.g. `ErrorEnvelope.error`). - `advisory` — explanatory text. Display or log it; **never switch on   it** (e.g. `SseWorkflowTerminalData.reason`). - `none` — not a vocabulary at all (an expression or an identifier,   e.g. `OptionSchema.pattern`). A field whose description hedges with \"common values\" or \"free-form\" must carry the marker; a test enforces it.  **Localisation (per ticket [I26](https://trello.com/c/rcnqwgI4)).**  Error responses + paused/blocked workflow statuses carry a localised human-readable `message` alongside a stable, never-localised `message_key`. Machine-readable fields (`error`, enum values, status codes) stay canonical English.  - **Currently committed locales:** `en-GB` only (per ticket   [`4GKyuYo6`](https://trello.com/c/4GKyuYo6)). The I26 carrier   shape (`Accept-Language` + `Content-Language` + `Vary` headers +   `locale` envelope field + `message_key` + `message_params`) is   stable and exercised; the **catalog** of translated `message`   strings is en-GB-only at runtime today. Additional locales (e.g.   `pt-PT`) will be advertised by name when their catalogs ship —   the request/response carrier shape does NOT change when a new   locale lands. Treat unrequested locales as \"machine-code +   `message_key` path is committed; localised `message` prose is   not\" until this prose enumerates them by name. - **Request:** `Accept-Language` header per RFC 9110 §12.5.4 (q-value   negotiation supported). The server selects the best-match locale   from its supported list; falls back to `en-GB` when no match —   which, until additional catalogs land, is every non-`en-GB`   `Accept-Language`. - **Response:** `Content-Language: <locale>` echo on every localised   response; `Vary: Accept-Language` on every response (CDN/cache   correctness — different `Accept-Language` requests produce   different responses). `Vary` is emitted unconditionally so the   header contract does not flip when a second locale ships. - **Fallback locale:** `en-GB` (also the canonical locale for   `message_key` translations and English `message` prose). - **SDK guidance:** switch on `error` (machine code) for typed   error branches; surface `message_key` to client-side i18n   catalogs (SDK companion work tracked at X19, cross-repo);   display `message` for end-user UI; **never parse `message` for   control flow** — it changes per locale.  Carrier shape lives on `ErrorEnvelope` (envelope-level optional `message_key` + `message` + `locale` + `message_params`) and `ValidationErrorEnvelope` (also per-`details[]` entry). Existing 402 / 403 / 422 envelopes (`BalanceExhaustedResponse`, `FeatureNotAvailableResponse`, `FeatureTierRestrictedResponse`, `WorkflowPausedDetail`) inherit the convention.  **Upload thresholds (per tickets [u0ar7Yye](https://trello.com/c/u0ar7Yye) + [58nBQLWQ](https://trello.com/c/58nBQLWQ)).** Canonical upload constants (single-shot cap, multipart chunk size, multipart concurrency default, multipart first-chunk size) live on the `UploadThresholds` schema with `const:`-pinned values. SDK generators emit these as typed binding constants so frontend / API / SDKs reference one source of truth instead of hardcoding magic numbers. A runtime `GET /api/uploads/limits` endpoint for dynamic discovery (per-tier / per-environment overrides) is a deferred follow-up.
  *
- * The version of the OpenAPI document: 2.217.0
+ * The version of the OpenAPI document: 2.219.0
  *
  *
  * NOTE: This class is auto generated by OpenAPI Generator (https://openapi-generator.tech).

package/dist/openapi/models/AccountLimits.d.ts

@@ -2,7 +2,7 @@
  * GISL Compression API
  * REST API for the GISL (Give It Smaller) file compression and processing service.  **Architecture:** - Upload files to get a `file_id` - Create workflows referencing uploaded files with operations (compress, thumbnail, image_watermark, text_watermark, merge, archive, convert, custom_luma, audio_overlay, audio_watermark) - Poll status, stream SSE events, or receive webhook callbacks - Download results per operation output  **Response envelope:** All mutation and query endpoints return `{ success: true, data: {...} }` on success and `{ success: false, error: \"...\", details: [...] }` on failure. Exceptions: `GET /api/operations/schema` returns raw JSON (per-tier private caching with ETag revalidation per ADR-0002 + I3), health probes return flat objects, and `POST /api/contact` returns 204 with no body.  **Availability metadata.** This spec uses the `x-availability` vendor extension as **decorative documentation only**. Per [ADR-0001](../docs/decisions/0001-contract-first-availability.md) §1.5, the runtime endpoint `GET /api/operations/schema` (ticket I3) is the authoritative source; the sidecar `availability.json` (ticket I3b) is the authoritative companion (generated, never hand-edited; CI cross-checks runtime ⇄ sidecar). SDKs MUST NOT depend on `x-availability` reaching generated code — code-generators that surface vendor extensions may emit it as documentation, but consumers read availability from the runtime endpoint, not from the generated bindings.  The 5-value vocabulary (`stable | beta | experimental | planned | deprecated`) is defined in the `AvailabilityValue` schema. See `schemas/FORMAT.md` §Availability Taxonomy for the operational rules (parser obligation: absent = stable; per-enum-value granularity is the `per_value_availability` primitive landed via ticket I17).  **Free-text string fields: `x-string-vocabulary` (ticket [`Q79yjcFF`](https://trello.com/c/Q79yjcFF)).** A `type: string` field with no `enum` that names example values says, as data, what a client may do with them (the same marker is used in the AsyncAPI document): - `open` — a vocabulary that grows. Each published value keeps its   meaning, the SET is not closed: switch on the values you know and   handle an unknown one as the generic case (e.g. `ErrorEnvelope.error`). - `advisory` — explanatory text. Display or log it; **never switch on   it** (e.g. `SseWorkflowTerminalData.reason`). - `none` — not a vocabulary at all (an expression or an identifier,   e.g. `OptionSchema.pattern`). A field whose description hedges with \"common values\" or \"free-form\" must carry the marker; a test enforces it.  **Localisation (per ticket [I26](https://trello.com/c/rcnqwgI4)).**  Error responses + paused/blocked workflow statuses carry a localised human-readable `message` alongside a stable, never-localised `message_key`. Machine-readable fields (`error`, enum values, status codes) stay canonical English.  - **Currently committed locales:** `en-GB` only (per ticket   [`4GKyuYo6`](https://trello.com/c/4GKyuYo6)). The I26 carrier   shape (`Accept-Language` + `Content-Language` + `Vary` headers +   `locale` envelope field + `message_key` + `message_params`) is   stable and exercised; the **catalog** of translated `message`   strings is en-GB-only at runtime today. Additional locales (e.g.   `pt-PT`) will be advertised by name when their catalogs ship —   the request/response carrier shape does NOT change when a new   locale lands. Treat unrequested locales as \"machine-code +   `message_key` path is committed; localised `message` prose is   not\" until this prose enumerates them by name. - **Request:** `Accept-Language` header per RFC 9110 §12.5.4 (q-value   negotiation supported). The server selects the best-match locale   from its supported list; falls back to `en-GB` when no match —   which, until additional catalogs land, is every non-`en-GB`   `Accept-Language`. - **Response:** `Content-Language: <locale>` echo on every localised   response; `Vary: Accept-Language` on every response (CDN/cache   correctness — different `Accept-Language` requests produce   different responses). `Vary` is emitted unconditionally so the   header contract does not flip when a second locale ships. - **Fallback locale:** `en-GB` (also the canonical locale for   `message_key` translations and English `message` prose). - **SDK guidance:** switch on `error` (machine code) for typed   error branches; surface `message_key` to client-side i18n   catalogs (SDK companion work tracked at X19, cross-repo);   display `message` for end-user UI; **never parse `message` for   control flow** — it changes per locale.  Carrier shape lives on `ErrorEnvelope` (envelope-level optional `message_key` + `message` + `locale` + `message_params`) and `ValidationErrorEnvelope` (also per-`details[]` entry). Existing 402 / 403 / 422 envelopes (`BalanceExhaustedResponse`, `FeatureNotAvailableResponse`, `FeatureTierRestrictedResponse`, `WorkflowPausedDetail`) inherit the convention.  **Upload thresholds (per tickets [u0ar7Yye](https://trello.com/c/u0ar7Yye) + [58nBQLWQ](https://trello.com/c/58nBQLWQ)).** Canonical upload constants (single-shot cap, multipart chunk size, multipart concurrency default, multipart first-chunk size) live on the `UploadThresholds` schema with `const:`-pinned values. SDK generators emit these as typed binding constants so frontend / API / SDKs reference one source of truth instead of hardcoding magic numbers. A runtime `GET /api/uploads/limits` endpoint for dynamic discovery (per-tier / per-environment overrides) is a deferred follow-up.
  *
- * The version of the OpenAPI document: 2.217.0
+ * The version of the OpenAPI document: 2.219.0
  *
  *
  * NOTE: This class is auto generated by OpenAPI Generator (https://openapi-generator.tech).

package/dist/openapi/models/AccountLimits.js

@@ -4,7 +4,7 @@
  * GISL Compression API
  * REST API for the GISL (Give It Smaller) file compression and processing service.  **Architecture:** - Upload files to get a `file_id` - Create workflows referencing uploaded files with operations (compress, thumbnail, image_watermark, text_watermark, merge, archive, convert, custom_luma, audio_overlay, audio_watermark) - Poll status, stream SSE events, or receive webhook callbacks - Download results per operation output  **Response envelope:** All mutation and query endpoints return `{ success: true, data: {...} }` on success and `{ success: false, error: \"...\", details: [...] }` on failure. Exceptions: `GET /api/operations/schema` returns raw JSON (per-tier private caching with ETag revalidation per ADR-0002 + I3), health probes return flat objects, and `POST /api/contact` returns 204 with no body.  **Availability metadata.** This spec uses the `x-availability` vendor extension as **decorative documentation only**. Per [ADR-0001](../docs/decisions/0001-contract-first-availability.md) §1.5, the runtime endpoint `GET /api/operations/schema` (ticket I3) is the authoritative source; the sidecar `availability.json` (ticket I3b) is the authoritative companion (generated, never hand-edited; CI cross-checks runtime ⇄ sidecar). SDKs MUST NOT depend on `x-availability` reaching generated code — code-generators that surface vendor extensions may emit it as documentation, but consumers read availability from the runtime endpoint, not from the generated bindings.  The 5-value vocabulary (`stable | beta | experimental | planned | deprecated`) is defined in the `AvailabilityValue` schema. See `schemas/FORMAT.md` §Availability Taxonomy for the operational rules (parser obligation: absent = stable; per-enum-value granularity is the `per_value_availability` primitive landed via ticket I17).  **Free-text string fields: `x-string-vocabulary` (ticket [`Q79yjcFF`](https://trello.com/c/Q79yjcFF)).** A `type: string` field with no `enum` that names example values says, as data, what a client may do with them (the same marker is used in the AsyncAPI document): - `open` — a vocabulary that grows. Each published value keeps its   meaning, the SET is not closed: switch on the values you know and   handle an unknown one as the generic case (e.g. `ErrorEnvelope.error`). - `advisory` — explanatory text. Display or log it; **never switch on   it** (e.g. `SseWorkflowTerminalData.reason`). - `none` — not a vocabulary at all (an expression or an identifier,   e.g. `OptionSchema.pattern`). A field whose description hedges with \"common values\" or \"free-form\" must carry the marker; a test enforces it.  **Localisation (per ticket [I26](https://trello.com/c/rcnqwgI4)).**  Error responses + paused/blocked workflow statuses carry a localised human-readable `message` alongside a stable, never-localised `message_key`. Machine-readable fields (`error`, enum values, status codes) stay canonical English.  - **Currently committed locales:** `en-GB` only (per ticket   [`4GKyuYo6`](https://trello.com/c/4GKyuYo6)). The I26 carrier   shape (`Accept-Language` + `Content-Language` + `Vary` headers +   `locale` envelope field + `message_key` + `message_params`) is   stable and exercised; the **catalog** of translated `message`   strings is en-GB-only at runtime today. Additional locales (e.g.   `pt-PT`) will be advertised by name when their catalogs ship —   the request/response carrier shape does NOT change when a new   locale lands. Treat unrequested locales as \"machine-code +   `message_key` path is committed; localised `message` prose is   not\" until this prose enumerates them by name. - **Request:** `Accept-Language` header per RFC 9110 §12.5.4 (q-value   negotiation supported). The server selects the best-match locale   from its supported list; falls back to `en-GB` when no match —   which, until additional catalogs land, is every non-`en-GB`   `Accept-Language`. - **Response:** `Content-Language: <locale>` echo on every localised   response; `Vary: Accept-Language` on every response (CDN/cache   correctness — different `Accept-Language` requests produce   different responses). `Vary` is emitted unconditionally so the   header contract does not flip when a second locale ships. - **Fallback locale:** `en-GB` (also the canonical locale for   `message_key` translations and English `message` prose). - **SDK guidance:** switch on `error` (machine code) for typed   error branches; surface `message_key` to client-side i18n   catalogs (SDK companion work tracked at X19, cross-repo);   display `message` for end-user UI; **never parse `message` for   control flow** — it changes per locale.  Carrier shape lives on `ErrorEnvelope` (envelope-level optional `message_key` + `message` + `locale` + `message_params`) and `ValidationErrorEnvelope` (also per-`details[]` entry). Existing 402 / 403 / 422 envelopes (`BalanceExhaustedResponse`, `FeatureNotAvailableResponse`, `FeatureTierRestrictedResponse`, `WorkflowPausedDetail`) inherit the convention.  **Upload thresholds (per tickets [u0ar7Yye](https://trello.com/c/u0ar7Yye) + [58nBQLWQ](https://trello.com/c/58nBQLWQ)).** Canonical upload constants (single-shot cap, multipart chunk size, multipart concurrency default, multipart first-chunk size) live on the `UploadThresholds` schema with `const:`-pinned values. SDK generators emit these as typed binding constants so frontend / API / SDKs reference one source of truth instead of hardcoding magic numbers. A runtime `GET /api/uploads/limits` endpoint for dynamic discovery (per-tier / per-environment overrides) is a deferred follow-up.
  *
- * The version of the OpenAPI document: 2.217.0
+ * The version of the OpenAPI document: 2.219.0
  *
  *
  * NOTE: This class is auto generated by OpenAPI Generator (https://openapi-generator.tech).