@warlock.js/ai 4.12.0 → 4.14.0

package/llms-full.txt

@@ -1626,7 +1626,7 @@ const gpt = openai.image({ name: "gpt-image-1", pricing: { input: 5, output: 40
 const dalle = openai.image({ name: "dall-e-3", pricing: { perImage: 0.04 } });
 ```
 
-A non-image model id (`openai.image({ name: "gpt-4o" })`) throws `InvalidRequestError` **at construction** — fail fast, like the embedder/vision guards.
+The model id is **not validated locally**. `openai.image({ name })` forwards the id to `client.images.generate` exactly as given, so a non-image id (`openai.image({ name: "gpt-4o" })`) constructs fine and fails at OpenAI — as a typed provider error on `result.error`, never as a local throw at construction.
 
 ## Google — Imagen (per-image)
 
@@ -1641,6 +1641,8 @@ const { data } = await ai.image({ model: imagen, prompt: "a watercolor lighthous
 
 Imagen returns base64 bytes (no hosted URL). When every candidate is safety-filtered, `ai.image` surfaces a typed `ContentFilterError` on `result.error`.
 
+Like every adapter, Google does **not** guard the model id: `google.image({ name })` passes the id through to `ai.models.generateImages` exactly as given, so an id Google does not serve fails as a typed provider error on `result.error` — not with a local throw at construction.
+
 ## Cost-truth — one rollup, two metering models
 
 `ai.image` fills `usage.cost` (a `ModelPricing`-shaped USD breakdown) so image spend folds into the **same** `Usage.cost` rollup as text — no second accounting path:
@@ -1713,7 +1715,7 @@ if (error) {
 }
 ```
 
-`SpeechModelContract` mirrors `EmbedderContract` / `ImageModelContract` — a peer primitive produced by the adapter's optional `speech?()` factory. An adapter without a TTS API simply doesn't define `speech()`, so calling it is a **compile-time** error, not a silent runtime failure. A non-TTS model id (`openai.speech({ name: "gpt-4o" })`) throws `InvalidRequestError` **at construction** — fail fast, like the embedder/image guards.
+`SpeechModelContract` mirrors `EmbedderContract` / `ImageModelContract` — a peer primitive produced by the adapter's optional `speech?()` factory. An adapter without a TTS API simply doesn't define `speech()`, so calling it is a **compile-time** error, not a silent runtime failure. The model id itself is **not validated locally** — a non-TTS id (`openai.speech({ name: "gpt-4o" })`) constructs fine and is forwarded to the provider as given, so it fails as a typed provider error on `result.error`, never as a local throw at construction.
 
 ## The result envelope
 
@@ -5150,7 +5152,7 @@ if (error) console.warn(error.code); // typed AIError
 else console.log(data.text);         // the transcript
 ```
 
-`TranscriptionModelContract` mirrors `SpeechModelContract` — a peer primitive produced by the adapter's optional `transcribe?()` factory. A non-STT model id (`openai.transcribe({ name: "gpt-4o" })`) throws `InvalidRequestError` **at construction** — fail fast, like the speech/embedder guards.
+`TranscriptionModelContract` mirrors `SpeechModelContract` — a peer primitive produced by the adapter's optional `transcribe?()` factory. The model id is **not validated locally** — a non-STT id (`openai.transcribe({ name: "gpt-4o" })`) constructs fine and is forwarded to the provider as given, so it fails as a typed provider error on `result.error`, never as a local throw at construction.
 
 ## The `AudioInput` shape + the two builders
 

package/package.json

@@ -15,16 +15,16 @@
     "@standard-schema/spec": "^1.0.0"
   },
   "peerDependencies": {
-    "@warlock.js/ai-openai": "4.12.0",
-    "@warlock.js/cache": "4.12.0",
-    "@warlock.js/logger": "4.12.0",
+    "@warlock.js/ai-openai": "4.14.0",
+    "@warlock.js/cache": "4.14.0",
+    "@warlock.js/logger": "4.14.0",
     "langfuse": "*",
     "openai": "*",
     "pdf-parse": "*",
     "pg": "*",
     "redis": "*"
   },
-  "version": "4.12.0",
+  "version": "4.14.0",
   "main": "./cjs/index.cjs",
   "module": "./esm/index.mjs",
   "types": "./esm/index.d.mts",

package/skills/generate-images/SKILL.md

@@ -81,7 +81,7 @@ const gpt = openai.image({ name: "gpt-image-1", pricing: { input: 5, output: 40
 const dalle = openai.image({ name: "dall-e-3", pricing: { perImage: 0.04 } });
 ```
 
-A non-image model id (`openai.image({ name: "gpt-4o" })`) throws `InvalidRequestError` **at construction** — fail fast, like the embedder/vision guards.
+The model id is **not validated locally**. `openai.image({ name })` forwards the id to `client.images.generate` exactly as given, so a non-image id (`openai.image({ name: "gpt-4o" })`) constructs fine and fails at OpenAI — as a typed provider error on `result.error`, never as a local throw at construction.
 
 ## Google — Imagen (per-image)
 
@@ -96,6 +96,8 @@ const { data } = await ai.image({ model: imagen, prompt: "a watercolor lighthous
 
 Imagen returns base64 bytes (no hosted URL). When every candidate is safety-filtered, `ai.image` surfaces a typed `ContentFilterError` on `result.error`.
 
+Like every adapter, Google does **not** guard the model id: `google.image({ name })` passes the id through to `ai.models.generateImages` exactly as given, so an id Google does not serve fails as a typed provider error on `result.error` — not with a local throw at construction.
+
 ## Cost-truth — one rollup, two metering models
 
 `ai.image` fills `usage.cost` (a `ModelPricing`-shaped USD breakdown) so image spend folds into the **same** `Usage.cost` rollup as text — no second accounting path:

package/skills/generate-speech/SKILL.md

@@ -26,7 +26,7 @@ if (error) {
 }
 ```
 
-`SpeechModelContract` mirrors `EmbedderContract` / `ImageModelContract` — a peer primitive produced by the adapter's optional `speech?()` factory. An adapter without a TTS API simply doesn't define `speech()`, so calling it is a **compile-time** error, not a silent runtime failure. A non-TTS model id (`openai.speech({ name: "gpt-4o" })`) throws `InvalidRequestError` **at construction** — fail fast, like the embedder/image guards.
+`SpeechModelContract` mirrors `EmbedderContract` / `ImageModelContract` — a peer primitive produced by the adapter's optional `speech?()` factory. An adapter without a TTS API simply doesn't define `speech()`, so calling it is a **compile-time** error, not a silent runtime failure. The model id itself is **not validated locally** — a non-TTS id (`openai.speech({ name: "gpt-4o" })`) constructs fine and is forwarded to the provider as given, so it fails as a typed provider error on `result.error`, never as a local throw at construction.
 
 ## The result envelope
 

package/skills/transcribe-audio/SKILL.md

@@ -33,7 +33,7 @@ if (error) console.warn(error.code); // typed AIError
 else console.log(data.text);         // the transcript
 ```
 
-`TranscriptionModelContract` mirrors `SpeechModelContract` — a peer primitive produced by the adapter's optional `transcribe?()` factory. A non-STT model id (`openai.transcribe({ name: "gpt-4o" })`) throws `InvalidRequestError` **at construction** — fail fast, like the speech/embedder guards.
+`TranscriptionModelContract` mirrors `SpeechModelContract` — a peer primitive produced by the adapter's optional `transcribe?()` factory. The model id is **not validated locally** — a non-STT id (`openai.transcribe({ name: "gpt-4o" })`) constructs fine and is forwarded to the provider as given, so it fails as a typed provider error on `result.error`, never as a local throw at construction.
 
 ## The `AudioInput` shape + the two builders