@fayegram-ai/decision-bridge 0.1.1 → 0.1.3

package/docs/guides/CLOUDFLARE_DECISIONS.md

@@ -16,8 +16,7 @@ short selectors and full forms.
 Questions and media consume context before text state; the service truncates
 remaining text state to fit. Requests whose questions/media exhaust context can
 fail. Byte checks cannot guarantee token fit or detect remote truncation.
-Clef's serving improvements and Clef-flash's lower pricing apply upstream without
-client configuration. Consult the official [model pages](#official-references)
+Consult the official [model pages](#official-references)
 for current pricing and availability; the bridge does not discover account access.
 
 ## Supported inference paths
@@ -108,7 +107,7 @@ extensions and the bridge's separate total deadline.
 
 ## Embedded images and limits
 
-Clef accepts up to four embedded PNG, JPEG, or WebP images. Native `images` entries
+Clef accepts embedded PNG, JPEG, or WebP images; the service owns count admission. Native `images` entries
 can be base64 data URLs or `{ content_type, base64 }` objects. Remote URLs are
 rejected. Common ordered input places all images before all text:
 
@@ -130,24 +129,15 @@ Text parts concatenate exactly, without inserted separators. Image detail can be
 omitted, `null`, or `auto`; explicit `low`, `high`, and `original` are unsupported.
 Safety identifiers are unsupported by this provider.
 
-| Boundary | Limit |
-| --- | --- |
-| Questions per request | 1–64 |
-| Question IDs | 1–100 ASCII letters, digits, `_`, `.`, or `-` |
-| Choice options | 2–255 |
-| Score levels | 2–10, ordered lowest first, indexed from zero |
-| Images | At most four |
-| Decoded image size | 4 MiB each; 8 MiB combined |
-| Image dimensions | Positive dimensions; at most 16 million pixels each |
-| Serialized request | Clef/Clef-flash: at most 13 MiB; Omni: 36 MiB by default, configurable through 64 MiB |
-
-Preflight checks supported formats, base64 syntax, headers, dimensions, and size.
-Header inspection is not full image decoding; the service can reject corrupt image data.
-Limits include execution options and binding envelope overhead. Omni's 36 MiB
-default is a bridge policy allowing base64 expansion of the permitted decoded
-media; its published schema specifies no whole-request byte ceiling. The effective
-request model selects the budget, including per-call overrides. An explicit
-`maxRequestBytes` remains a caller ceiling and is never enlarged by selecting Omni.
+Preflight validates supported representations, canonical base64, positive image
+dimensions and container structure. Service question/media counts, decoded sizes,
+dimensions, duration and context ceilings are not enforced locally. Header
+inspection is not full decoding; upstream admission and decoding remain authoritative.
+
+Client request budgets default to 13 MiB for Clef/Clef-flash and 36 MiB for Omni.
+An explicit `maxRequestBytes` overrides either default through the shared 64 MiB
+client ceiling. Binding envelope overhead counts toward that budget. Response
+budgets remain separate. No automatic resizing, splitting or truncation occurs.
 
 Cloudflare's four-decimal probability precision receives a count-aware sum
 allowance: the larger of 0.001 and half a decimal unit per option, plus floating-point
@@ -204,19 +194,10 @@ audio/video parts accept only `type` and `dataUrl`; remote URLs, paths, image
 detail, and caller-provided durations are rejected. Other providers and the older
 Clef models reject audio/video before dispatch.
 
-| Media | Count | Decoded size per item | Duration per item |
-| --- | ---: | ---: | ---: |
-| Audio | 4 | 8 MiB | 300 seconds |
-| Video | 2 | 16 MiB | 60 seconds |
-
-Audio and video together are limited to 16 MiB decoded. Image limits above remain
-separate. Bounded inspectors validate canonical base64, MIME/header consistency,
-and container structure, and reject excessive duration when PCM/float WAV,
-ordinary MP3 frames, MP4 metadata, or WebM metadata establishes it reliably.
-Full decoding, codec validity, and duration that cannot be established locally
-(including compressed WAV, free-format MP3, or missing container duration) remain
-service validation. The bridge performs no transcription, transcoding, frame
-extraction, or streaming, and adds no runtime media dependency.
+Inspectors validate canonical base64, MIME/header consistency and supported
+container structure. Service count, decoded-size and duration limits are handled
+as upstream errors. The bridge performs no transcription, transcoding, frame
+extraction or streaming, and adds no runtime media dependency.
 
 Cloudflare samples video at two frames per second. Video soundtracks are processed
 only when **every supplied video has a soundtrack**. Common results use
@@ -302,23 +283,20 @@ binding sends Gateway settings in the third `ai.run()` argument.
 
 | Setting | Accepted value |
 | --- | --- |
-| `id` | Required nonblank printable ASCII string, 1–64 characters |
-| `cacheKey` | Optional nonblank printable ASCII string, 1–200 characters |
-| `cacheTtl` | Integer seconds, 60–2,592,000 (30 days) |
+| `id` | Required nonblank printable ASCII string |
+| `cacheKey` | Optional nonblank printable ASCII string |
+| `cacheTtl` | Nonnegative integer seconds; service cache policy applies upstream |
 | `skipCache`, `collectLog` | Optional booleans |
-| `metadata` | At most five entries with JSON primitive values; caller-supplied `cf.*` keys are reserved |
+| `metadata` | JSON primitive values; caller-supplied `cf.*` keys are reserved |
 | `requestTimeoutMs` | Positive integer milliseconds, up to 2,147,483,647 |
-| `retries.maxAttempts` | 1–5; defaults to 1 when Gateway is configured |
-| `retries.retryDelayMs` | 0–60,000 milliseconds |
+| `retries.maxAttempts` | Positive integer; defaults to 1 when Gateway is configured |
+| `retries.retryDelayMs` | Nonnegative integer milliseconds |
 | `retries.backoff` | `constant`, `linear`, or `exponential` |
 
 The shared validator rejects invalid Gateway settings with `CONFIGURATION_ERROR`
-before dispatch. Metadata is preserved unchanged: the bridge rejects entries
-that [Cloudflare would ignore or remove](https://developers.cloudflare.com/ai-gateway/observability/custom-metadata/)
-instead of silently dropping them. Cloudflare's [cache TTL documentation](https://developers.cloudflare.com/ai-gateway/features/caching/)
-describes the upper limit as one month; its [official numeric reference](https://github.com/cloudflare/skills/blob/0cb6116228ad153126f474dce1dfbd05f7944604/skills/cloudflare/references/ai-gateway/README.md)
-specifies 2,592,000 seconds (30 days). Cache TTL and first-response timeout have
-separate validation bounds.
+before dispatch. Metadata is preserved unchanged. Reserved `cf.*` keys and
+non-primitive values are rejected; service entry-count and cache admission
+limits remain upstream. The adapter does not infer Gateway acceptance.
 
 Gateway timeout covers time to the first response; the bridge's total deadline
 continues to bound the call. Gateway logging/caching follows the explicit options

package/docs/guides/CUSTOMIZATION.md

@@ -1,6 +1,7 @@
 # Customizing decisions and extending the bridge
 
-Customization is **per request**. You supply context in `state`, the judgment in
+Customization is **per request**. You supply context in `state`, ordered `input`
+or `conversation` according to provider capabilities, the judgment in
 `instructions`, and the allowed categories/rubric in `criteria`. The same bridge
 instance can evaluate unrelated tasks on successive calls. This changes the
 question sent to the existing model, not its weights or account configuration.
@@ -79,6 +80,10 @@ const bridge = createBridge({
 Extensions run in registration order. They receive immutable snapshots and must
 return replacements rather than mutate input. Async retrieval can use the supplied
 context signal. No content is persisted unless your extension explicitly saves it.
+Use `prepareInput(parts, context)` for ordered input and
+`prepareConversation(items, context)` for conversation history. A preparation
+extension must implement the selected form's hook; transformed content is
+revalidated before dispatch. Observer-only extensions work with every form.
 
 ## Observe without changing decisions
 
@@ -133,7 +138,8 @@ and [SDK validation contract](../reference/SDK.md#additional-exports).
 
 ## Develop additions without changing the public call
 
-Add retrieval/redaction through `prepareState` or `prepareInput`, metrics through
+Add retrieval/redaction through `prepareState`, `prepareInput` or
+`prepareConversation`, metrics through
 `onResult`, and another protocol through the explicit [custom decision](CUSTOM_DECISION.md)
 or [custom generative](CUSTOM_GENERATIVE.md) contract. Keep routing/threshold policies
 outside the provider and cover additions with relevant regression tests. Persisted templates,

package/docs/guides/CUSTOM_GENERATIVE.md

@@ -72,8 +72,9 @@ or zero-based level indices. It checks probabilities, derives argmax/weighted
 scores and matching legends, and marks absent usage as unknown. It does not repair
 JSON, strip Markdown, or create confidence. `buildPrompt` and `normalizeGenerated`
 independently validate and snapshot the request. They require shared state and
-legacy questions, and reject ordered input, typed choices, labelled scores, and
-safety identifiers with `UNSUPPORTED_CAPABILITY`. Malformed request JSON raises
+legacy questions, and reject ordered input, conversation/tool history, remote
+images, prompt-cache controls, typed choices, labelled scores and safety
+identifiers with `UNSUPPORTED_CAPABILITY`. Malformed request JSON raises
 `VALIDATION_ERROR`; unsupported input is never silently discarded. New modules
 implement their own encoding and normalization for richer capabilities.
 
@@ -110,7 +111,8 @@ provenance and `confidenceSource: 'unavailable'`, and rejects returned native
 confidence claims or answer confidence fields. Their output is an estimate elicited
 from a model, not a native probability, logprob, or calibration guarantee.
 Presets support shared state and legacy `choice`, `noul`, and `score`; ordered
-media, typed choices, labelled scores, and safety identifiers are unsupported.
+media, conversation/tool history, remote images, prompt-cache controls, typed
+choices, labelled scores and safety identifiers are unsupported.
 A new module may declare capabilities only after implementing and validating them.
 
 ## OpenAI-compatible servers

package/docs/guides/CUSTOM_PROVIDERS.md

@@ -41,7 +41,9 @@ Internal transport helpers are not stable public exports. Delegate a maintained
 preset for a protocol already supported, or implement and test your own transport.
 
 Capability omissions retain the legacy shared-state and question profile. Audio
-and video additions default to false/zero. Declare richer forms only when their
+and video additions default to false/zero; conversation, tool history, remote
+images and prompt caching default to false. Explicit quantitative custom limits
+are enforced; `null` leaves that admission limit to the service. Declare richer forms only when their
 encoding and validation are implemented. The bridge runs preflight again after
 preparation hooks. Unknown usage is null; no provider should invent token counts,
 confidence, or calibration. Generative native-provenance claims are rejected.
@@ -68,8 +70,8 @@ relative to that file; flag paths
 resolve relative to the working directory. Explicit modules execute trusted code;
 this extension point is not a sandbox. No module is discovered automatically.
 
-`init` creates an honest service implementation placeholder when no implementation
-is selected. When given an existing module, it references that module without
+`init` creates a provider module with an unimplemented service call when no
+implementation is selected. When given an existing module, it references that module without
 loading or overwriting it. Offline `validate` never imports custom modules or
 requires credentials; it reports structure-only validation for unknown protocols.
 `bridge.validate()` on an instantiated provider includes capabilities and preflight.

package/docs/guides/EXTRACTION.md

@@ -0,0 +1,125 @@
+# Clairic extraction
+
+Requires **0.1.3 or later**. See [version availability](../README.md#version-availability)
+and [installation](GETTING_STARTED.md#install-into-your-application).
+
+Import extraction from `@fayegram-ai/decision-bridge/extraction`. This operation
+returns typed values and predictive statistics, using Cloudflare Clairic through
+REST or an injected Workers AI binding. Its result is independent of the decision
+bridge's question/answer contract.
+
+```javascript
+import { CloudflareExtractionProvider } from '@fayegram-ai/decision-bridge/extraction';
+
+const provider = new CloudflareExtractionProvider();
+const { extraction, meta } = await provider.extract({
+    state: 'Invoice total: 1080 USD. Due date: 2026-11-01.',
+    reference_date: '2026-10-20',
+    fields: {
+        amount: { type: 'number', unit: 'USD' },
+        due_date: { type: 'date' },
+        payment_days: { type: 'estimate', min: 0, max: 90, step: 1 },
+    },
+    options: { include_quantiles: true, return_state_text: true },
+});
+console.log(extraction.answers.amount.value, extraction.truncated, meta.requestId);
+```
+
+REST reads `CLOUDFLARE_API_TOKEN` and `CLOUDFLARE_ACCOUNT_ID` only for the official
+origin, or accepts explicit `apiKey` and `accountId`. Custom origins require
+explicit credentials. The default model is `clairic`; the full selector
+`@cf/cloudflare/clairic` is also accepted. Requested, resolved and reported model
+identities remain separate. There is no model-discovery request.
+
+## Fields and results
+
+| Type | Definition | Returned value |
+| --- | --- | --- |
+| `text` | Instructions describing a text value | String or null |
+| `number` | Optional unit, integer and decimal settings | Number or null |
+| `date`, `time`, `datetime` | Instructions describing the temporal value | String or null; no implicit timezone conversion |
+| `duration` | Unit: seconds, minutes, hours, days, weeks, months or years | Number or null |
+| `list` | Required `items`: text, number or date | Ordered list, including an empty list, or null |
+| `estimate` | Required numeric or ISO-date `min` and `max`; optional `step` and `scale` | Number, date string or null, plus available distribution summaries |
+| `forecast` | Prediction bounds plus optional positive `horizon` and `interval` | Native value and ordered predictive `steps`, retained as JSON |
+
+`instructions` and `null_threshold` describe individual fields. `null_threshold`
+must be greater than zero and at most one. `number` and `duration` accept boolean
+`integer` and nonnegative integer `decimals`; numeric lists also accept `decimals`.
+Grids support positive numeric steps or `month`/`year`;
+`scale` is `linear` or `log`. Forecast intervals are `step`, `minute`, `hour`,
+`day`, `week`, `month`, `quarter`, or `year`. Server ceilings remain server-owned.
+
+The immutable envelope is `{ extraction, meta }`. The native extraction retains
+all returned fields: values, guesses, confidence, `p_null`, spans, means,
+medians, `p10`/`p90`, bounds probabilities, quantiles, forecast steps, `truncated`,
+optional `state_text`, and usage including optional pass counts. Unknown native
+response fields are retained after safe JSON validation.
+
+A null value can coexist with a guess and span. Typed output does not establish
+factual accuracy. The adapter does not recalculate predictions or token totals.
+Predictive substructures with an underspecified schema remain JSON values in the
+types; span offsets are preserved without assuming JavaScript slicing units.
+
+## Images, binding and Gateway
+
+Provide text `state`, embedded `images`, or both. Images accept PNG/JPEG/WebP data
+URLs or `{ content_type, base64 }` objects. Remote images are not supported by
+this operation. The adapter checks format and container structure, while the
+service owns image-count, decoded-size, dimension and context admission.
+
+```javascript
+const provider = new CloudflareExtractionProvider({
+    transport: 'binding',
+    ai: env.AI,
+    gateway: { id: 'existing-gateway', skipCache: true },
+});
+```
+
+`include_quantiles` and `return_state_text` stay in the model input's `options`.
+`rejectIfBusy` is an execution setting, configurable on the provider or request
+options; binding dispatch places it in the third `ai.run` argument. Gateway
+settings use the same routing and diagnostics as the Clef provider. A service
+rejection of an execution setting is reported without dropping the setting.
+
+## Validation, batching and errors
+
+`validateExtractionRequest(request, options?)` is credential-free offline
+validation. `provider.validate(request)` uses the configured provider settings.
+Both check structure and client resource budgets, without promising admission.
+
+Request/response budgets default to 8 MiB each and can be configured through
+64 MiB. Requests are checked before media decoding. `extract(request, { signal })`
+accepts an `AbortSignal`. Set `timeoutMs` on the provider constructor to bound
+validation, dispatch and response handling; extraction call options accept only
+`signal` and have no per-call timeout override.
+Retries default to zero. Billing/allocation errors are not retried even when
+retries are enabled; temporary eligible failures use the shared error policy.
+See [errors and recovery](../reference/ERRORS.md).
+
+`extractBatch(provider, requests, { concurrency, signal, stopOnError })` validates
+all array requests before dispatch and returns ordered `{ index, ok, result }`
+or `{ index, ok, error }` records. Concurrency defaults to four and is a client
+execution setting, independent of any service quota.
+Batch results retain each caller's `requestedModel`, including the full
+`@cf/cloudflare/clairic` alias, while `resolvedModel` and dispatched input use
+`clairic`, just as for individual extraction calls.
+
+## CLI
+
+```shell
+decision extraction init my-extraction
+decision extraction validate request.json
+decision extraction run request.json --env-file .env
+decision extraction batch requests.jsonl --concurrency 4 --env-file .env
+```
+
+The namespace defaults to `cloudflare-extraction`; incompatible explicit
+providers are rejected. Configuration schema version remains 1. Use the existing
+Cloudflare account, credential, Gateway, timeout and retry settings. Inference
+options belong in request JSON. Validation is offline and reports `fieldCount`.
+Streaming JSONL batches can complete valid rows before a later row fails.
+
+See the [runnable extraction example](../../examples/extraction.mjs), the
+[official model schema](https://developers.cloudflare.com/workers-ai/models/clairic/)
+and [extraction concepts](https://developers.cloudflare.com/workers-ai/concepts/extraction-models/).

package/docs/guides/GETTING_STARTED.md

@@ -20,18 +20,22 @@ In your application's directory, install the package from npm:
 npm install @fayegram-ai/decision-bridge
 ```
 
-You can also install a verified local archive, replacing the path below. See
-[packaging and release](../development/RELEASING.md) to build and verify one from
-a source checkout:
+You can also install a verified package archive. Replace the example archive
+path with your own; see [packaging and release](../development/RELEASING.md) to
+build and verify one from a source checkout:
 
 ```bash
-npm install /path/to/fayegram-ai-decision-bridge-0.1.1.tgz
+npm install /path/to/fayegram-ai-decision-bridge-0.1.3.tgz
 ```
 
 The installed package contains the SDK, declarations, command runner, examples,
 and docs. It runs without compilation; source development scripts and tests are
 excluded.
 
+Check the installed version with `npm exec -- decision-bridge --version`.
+Use the documentation bundled with your installed package, or select the
+matching [versioned documentation](../README.md#version-availability).
+
 ## Configure credentials
 
 Create `.env` in your application's directory and edit it locally:

package/docs/guides/JEV.md

@@ -0,0 +1,167 @@
+# Jev images, boxes, and long context
+
+Jev uses TypeSafe's native `POST /v1/systemone` endpoint. The SDK defaults to
+`jev-latest`; `jev-preview` and explicit version IDs are also accepted. Model
+listing is account-specific and may list aliases without versioned IDs.
+The returned `model` records the service-reported version independently of
+`meta.requestedModel`.
+
+## Experimental opt-ins
+
+Jev images and long context are disabled by default. Enable each independently
+in provider configuration. TypeSafe treats these capabilities as experimental;
+their accuracy, formats, and limits can change.
+
+```javascript
+import { createBridge, JevProvider, box, noul } from '@fayegram-ai/decision-bridge';
+
+const provider = new JevProvider({
+    model: 'jev-1.14.0',
+    dangerZone: { images: true, longContext: true },
+});
+const bridge = createBridge({ provider });
+
+// imageBase64 is a caller-prepared PNG or JPEG, without a data URL prefix.
+const result = await bridge.evaluate({
+    input: [
+        { type: 'text', text: 'A photograph from the kitchen camera.' },
+        { type: 'image', dataUrl: 'data:image/png;base64,' + imageBase64 },
+    ],
+    questions: {
+        present: noul('Is there a coffee mug in the image?'),
+        mug: box('The coffee mug in the image'),
+    },
+});
+console.log(result.answers.present.noul, result.answers.mug.box);
+```
+
+`dangerZone.images` sends `TypeSafe-Danger-Zone-Images: true` and enables common
+image input and boxes. It works with Jev 1.13 and 1.14 for images. Boxes require
+`jev-preview` or a pinned Jev version of 1.14.0 or newer. Local box capability
+lookup remains conservative for moving stable aliases and unknown model names;
+use a supported pinned ID or `jev-preview` for boxes. Account access and hosted
+model availability are still checked by the service.
+
+`dangerZone.longContext` sends `TypeSafe-Danger-Zone-Long-Context: true`. TypeSafe
+documents 262k tokens for both the per-request and per-question contexts with
+this header, compared with default limits of 64k and 32k respectively. This
+option works on Jev 1.13 and 1.14. It does not change the byte budget or estimate
+token counts. Longer contexts are slower and less accurate.
+
+Ordered common input becomes a state array in the same order. Text stays text;
+each image becomes an asset reference. The SDK chooses references that do not
+collide with the input or question text. Jev does not support image detail
+controls: omit `detail`, or use `null` or `auto`. Audio/video, typed choices,
+labelled scores, and safety identifiers remain unsupported by this provider.
+
+## Native assets in instructions and criteria
+
+Use `provider.createDecision()` when images belong inside structured state,
+instructions, or criteria. It accepts the native JSON request shape and returns
+`{ decision, meta }`. Native calls share validation, transport, cancellation,
+and retry policy with common calls, and bypass bridge extensions and observers.
+
+```javascript
+const { decision, meta } = await provider.createDecision({
+    state: { photo: 'photo-unique-reference' },
+    assets: {
+        'photo-unique-reference': {
+            type: 'image', format: 'png', data_base64: imageBase64,
+        },
+        'example-unique-reference': {
+            type: 'image', format: 'jpg', data_base64: exampleBase64,
+        },
+    },
+    questions: {
+        same: noul('Does the object in `photo` match this picture: example-unique-reference?'),
+        target: box('The matching object in `photo`'),
+    },
+});
+console.log(decision.answers.target.box, meta.requestedModel);
+```
+
+References are substituted in string values, including inside sentences and
+nested JSON. Object keys are not substituted. Every asset must be referenced.
+Whole-value references are resolved first. In longer strings, matches are
+consumed from left to right; when keys overlap at the same position, the current
+API uses asset-map order. A reference hidden by another match does not count as
+used. Preflight rejects unused assets before dispatch. Place references in
+separate values when both images are intended.
+Use long unique references (for example, fresh UUIDs) that cannot accidentally
+match ordinary text. Each reference is at most 1024 UTF-8 bytes; all reference
+keys together are at most 65536 bytes. A file is uploaded once, but repeating
+its reference repeats the image in the model's expanded input.
+
+The bridge uses the JSON/base64 transport supported by TypeSafe. It does not
+accept multipart `part` assets, read local image files, fetch URLs, resize,
+rotate, or re-encode images. Base64 overhead counts toward the request limit.
+Native Jev assets are separate from the common request shape; use
+`createDecision` for images embedded in instructions or criteria.
+
+## Boxes and results
+
+`box(instructions)` takes no criteria. Its answer has `type: 'box'`, `box`,
+and native `confidence`. Coordinates are fractions measured from the top-left:
+`left` and `right` scale with width; `top` and `bottom` scale with height.
+All four values must be finite numbers in 0–1, with `left < right` and
+`top < bottom`. Validation rejects malformed boxes rather than repairing them.
+Common results containing box questions use result `meta.schemaVersion: 4`;
+the CLI configuration schema version remains 1.
+
+Each question returns one box. There is no absent-target result: a box is
+returned even when the target is absent. Pair it with a presence `noul` and
+choose your own presence/confidence thresholds using representative images.
+Jev token usage may be `null`; that means unknown, never zero. Native confidence
+and all other result checks remain enforced.
+
+## Images and client budgets
+
+PNG/JPEG formats, canonical base64, positive dimensions, still-image structure
+and asset reference semantics are checked locally. Service image dimensions,
+decoded/processed sizes and context admission remain upstream. The bridge does
+not resize, rotate or re-encode images. Prepare upright pixels in the application.
+
+Jev requests default to a 32 MiB client serialized-JSON budget. An explicit
+`maxRequestBytes` can select a budget through 64 MiB. Responses default to 8 MiB.
+General JSON processing retains a 100-level depth and 200,000-node safeguard;
+reference-pattern safeguards bound local regex construction. These are client
+resource controls, independent of service quotas.
+
+## CLI
+
+```json
+{
+    "schemaVersion": 1,
+    "provider": "jev",
+    "model": "jev-1.14.0",
+    "apiKeyEnv": "TYPESAFE_API_KEY",
+    "dangerZone": { "images": true, "longContext": true }
+}
+```
+
+```shell
+decision-bridge validate request.json --config decision.config.json
+decision-bridge evaluate request.json --config decision.config.json --env-file .env
+decision-bridge validate native.json --native --model jev-preview --jev-images
+decision-bridge batch native.jsonl --native --config decision.config.json --env-file .env
+```
+
+Common/native validation is offline and uses the same preflight as inference.
+`--native` works with `init`, `validate`, `evaluate`, and `batch`. Jev keeps its
+`models` command. CLI JSON/JSONL input defaults to 32 MiB per record and permits a larger explicit client budget through 64 MiB;
+the actual serialized request must also fit the configured provider ceiling.
+
+`--jev-images` / `--no-jev-images` and `--jev-long-context` /
+`--no-jev-long-context` override independent fields in `dangerZone`.
+`DECISION_JEV_IMAGES` and `DECISION_JEV_LONG_CONTEXT` accept exactly `true` or
+`false`. Flags override environment, then JSON configuration, then defaults.
+Switching providers discards Jev-specific settings; using these flags with
+another provider is a configuration error.
+
+## Protocol sources
+
+- [TypeSafe release announcement](https://typesafe.ai/blog/jev-1-14)
+- [Images, assets, limits, and experimental header](https://docs.typesafe.ai/images)
+- [Box questions and absence behavior](https://docs.typesafe.ai/primitives/box)
+- [Models, aliases, and context limits](https://docs.typesafe.ai/models)
+- [JavaScript SDK compatibility changes](https://docs.typesafe.ai/sdk/javascript/changelog)