@ai-sdk/xai 4.0.0-canary.73 → 4.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,158 @@
1
1
  # @ai-sdk/xai
2
2
 
3
+ ## 4.0.0
4
+
5
+ ### Major Changes
6
+
7
+ - f7d4f01: feat(provider): add support for `reasoning-file` type for files that are part of reasoning
8
+ - 776b617: feat(provider): adding new 'custom' content type
9
+ - ef992f8: Remove CommonJS exports from all packages. All packages are now ESM-only (`"type": "module"`). Consumers using `require()` must switch to ESM `import` syntax.
10
+ - c29a26f: feat(provider): add support for provider references and uploading files as supported per provider
11
+ - 8359612: Start v7 pre-release
12
+ - 04e9009: chore: make provider implementations code patterns more consistent, including renaming certain exported symbols
13
+
14
+ For all externally exported symbols that were renamed, the old names continue to work via deprecated aliases.
15
+
16
+ - f62681f: feat(provider/xai): make responses api the default
17
+
18
+ ### Patch Changes
19
+
20
+ - 38fc777: Add AI Gateway hint to provider READMEs
21
+ - 85735d8: fix(xai): stop emitting additionalProperties flag
22
+ - 74d520f: feat: migrate providers to support new top-level `reasoning` parameter
23
+ - 25f1837: feat(xai): add b64_json response format, usage cost tracking, and quality/user parameters for image models
24
+ - a0b0a0c: expose costInUsdTicks in responses provider metadata
25
+ - aa5a583: chore(provider/xai): remove obsolete Grok 2 models now that they are shut down in their API
26
+ - e5bdc8d: fix (provider/xai): handle mid-stream error chunks
27
+ - 4799fa5: chore(provider/xai): update provider to use v4 types
28
+ - 813851f: fix (provider/xai): add response.incomplete and response.failed streaming event handling
29
+ - 2dc2a52: fix reasoning-start dedupe on multi-summary-part responses to prevent xai 400 on continuation requests
30
+ - 9f0e36c: trigger release for all packages after provenance setup
31
+ - 58a2ad7: fix: more precise default message for tool execution denial
32
+ - d20829e: feat(provider/xai): add moderation error, and costInUsdTicks to video model
33
+ - 7fc6bd6: Raise minimum supported Node.js version to 22. Supported versions: 22, 24, and 26.
34
+ - 0f11f10: add reasoningSummary to responses API provider options
35
+ - 23f9d72: deprecate searchParameters (xai live search) in favor of web_search/x_search agent tools
36
+ - 0c4c275: trigger initial canary release
37
+ - 1293885: surface full xai error detail in APICallError.message instead of falling back to http status text
38
+ - 8d87577: fix(xai): support encrypted reasoning round-trip for ZDR
39
+ - 9f20868: fix(provider/xai): correct finish reason for tool calls
40
+ - 4f91b5d: chore(provider/xai): update Grok 4.20 model IDs to their non-beta versions
41
+ - ce769dd: feat(provider): add experimental Realtime API support for voice conversations
42
+
43
+ Adds first-class support for realtime (speech-to-speech) APIs:
44
+
45
+ - `Experimental_RealtimeModelV4` spec in `@ai-sdk/provider` with normalized event types and factory
46
+ - OpenAI, Google, and xAI realtime provider implementations
47
+ - `openai.experimental_realtime()` / `google.experimental_realtime()` / `xai.experimental_realtime()` work in both server and browser
48
+ - `.getToken()` static method on each provider for server-side ephemeral token creation
49
+ - `experimental_getRealtimeToolDefinitions` helper for provider session tool definitions
50
+ - `experimental_useRealtime` hook in `@ai-sdk/react` returning `UIMessage[]` (aligned with `useChat`), with `onToolCall` and `addToolOutput` for client-driven tool execution
51
+ - `inputAudioTranscription` session config for showing transcribed user audio messages when supported by the provider
52
+
53
+ - 12115e9: fix reasoning text extraction from content in responses doGenerate
54
+ - 9bd6512: feat(provider): change file part data property to be tagged with a type and remove the image part type
55
+ - 258c093: chore: ensure consistent import handling and avoid import duplicates or cycles
56
+ - b8396f0: trigger initial beta release
57
+ - f51c95e: feat(provider/xai): add video extension and reference-to-video (R2V) support
58
+ - 90e2d8a: chore: fix unused vars not being flagged by our lint tooling
59
+ - f5181ad: feat(provider/xai): support multiple input images for image editing
60
+ - b3976a2: Add workflow serialization support to all provider models.
61
+
62
+ **`@ai-sdk/provider-utils`:** New `serializeModel()` helper that extracts only serializable properties from a model instance, filtering out functions and objects containing functions. Third-party provider authors can use this to add workflow support to their own models.
63
+
64
+ **All providers:** `headers` is now optional in provider config types. This is non-breaking — existing code that passes `headers` continues to work. Custom provider implementations that construct model configs manually can now omit `headers`, which is useful when models are deserialized from a workflow step boundary where auth is provided separately.
65
+
66
+ All provider model classes now include `WORKFLOW_SERIALIZE` and `WORKFLOW_DESERIALIZE` static methods, enabling them to cross workflow step boundaries without serialization errors.
67
+
68
+ - 1dbecd7: feat(provider/xai): add `enableImageSearch` to the xAI Web Search tool
69
+
70
+ The xAI Responses API supports `enable_image_search` on Web Search tools. `xai.tools.webSearch()` now accepts `enableImageSearch` and sends it through to the API as `enable_image_search`.
71
+
72
+ - 80e1702: feat(xai): support `'none'` and `'medium'` reasoning effort for `grok-4.3`,
73
+ and curate the model ID autocomplete list
74
+
75
+ `grok-4.3` accepts `reasoning_effort` of `"none" | "low" | "medium" | "high"`,
76
+ where `"none"` disables reasoning entirely (no thinking tokens) and `"medium"`
77
+ provides more thinking for less-latency-sensitive applications.
78
+
79
+ - Adds `'none'` to the allowed values for `providerOptions.xai.reasoningEffort`
80
+ on both the chat (`xai()`) and responses (`xai.responses()`) language models.
81
+ - Adds `'medium'` to the chat model's `reasoningEffort` enum (the responses
82
+ model already supported it).
83
+ - Top-level `reasoning: 'medium'` now maps to `reasoning_effort: 'medium'` for
84
+ the chat model (previously it was coerced to `'low'` because `'medium'` was
85
+ not a valid value).
86
+
87
+ In addition, the `XaiChatModelId` and `XaiResponsesModelId` autocomplete unions
88
+ have been trimmed to xAI's current model lineup
89
+ ([docs](https://docs.x.ai/docs/models)):
90
+
91
+ - `grok-4.20-non-reasoning`
92
+ - `grok-4.20-reasoning`
93
+ - `grok-4.3`
94
+ - `grok-latest`
95
+
96
+ Older entries (`grok-3*`, `grok-4`, `grok-4-0709`, `grok-4-latest`,
97
+ `grok-4-1-fast-*`, `grok-4-fast-*`, `grok-code-fast-1`, and
98
+ `grok-4.20-multi-agent-0309`) have been removed from the autocomplete list.
99
+ This is **not** a runtime change — the model ID type is still open
100
+ (`(string & {})`), so passing any model ID that the xAI API accepts continues
101
+ to work; only IDE autocomplete is affected.
102
+
103
+ ```ts
104
+ import { xai } from "@ai-sdk/xai";
105
+ import { generateText } from "ai";
106
+
107
+ await generateText({
108
+ model: xai("grok-4.3"),
109
+ prompt: "Hi",
110
+ providerOptions: {
111
+ xai: { reasoningEffort: "none" },
112
+ },
113
+ });
114
+ ```
115
+
116
+ - 78b6433: feat(provider/xai): support non-image file parts (PDF, text, CSV) in the Responses API via `input_file` + `file_url`
117
+
118
+ The xAI Responses API accepts `{ type: 'input_file', file_url }` for non-image documents (see https://docs.x.ai/docs/guides/chat-with-files), but the AI SDK xAI Responses provider previously threw `UnsupportedFunctionalityError` for any file part whose `mediaType` did not start with `image/`.
119
+
120
+ When a file part is passed with `data: URL` and a non-image media type, the provider now emits `{ type: 'input_file', file_url }`. `application/pdf` and `text/*` are also added to `supportedUrls` so the SDK does not download them to bytes before reaching the converter.
121
+
122
+ Inline-byte (base64) inputs for non-image media types continue to throw, since xAI's Responses API requires either a public URL or a pre-uploaded `file_id` for non-image documents.
123
+
124
+ - 7486744: Add xAI speech-to-text transcription support.
125
+ - 7486744: feat(provider/xai): add text-to-speech support
126
+
127
+ ## 4.0.0-beta.76
128
+
129
+ ### Patch Changes
130
+
131
+ - Updated dependencies [0416e3e]
132
+ - @ai-sdk/provider@4.0.0-beta.20
133
+ - @ai-sdk/openai-compatible@3.0.0-beta.58
134
+ - @ai-sdk/provider-utils@5.0.0-beta.50
135
+
136
+ ## 4.0.0-beta.75
137
+
138
+ ### Patch Changes
139
+
140
+ - b8396f0: trigger initial beta release
141
+ - Updated dependencies [b8396f0]
142
+ - @ai-sdk/openai-compatible@3.0.0-beta.57
143
+ - @ai-sdk/provider-utils@5.0.0-beta.49
144
+ - @ai-sdk/provider@4.0.0-beta.19
145
+
146
+ ## 4.0.0-canary.74
147
+
148
+ ### Patch Changes
149
+
150
+ - Updated dependencies [aeda373]
151
+ - Updated dependencies [375fdd7]
152
+ - Updated dependencies [b4507d5]
153
+ - @ai-sdk/provider-utils@5.0.0-canary.48
154
+ - @ai-sdk/openai-compatible@3.0.0-canary.56
155
+
3
156
  ## 4.0.0-canary.73
4
157
 
5
158
  ### Patch Changes
package/dist/index.js CHANGED
@@ -3393,7 +3393,7 @@ var xaiTools = {
3393
3393
  };
3394
3394
 
3395
3395
  // src/version.ts
3396
- var VERSION = true ? "4.0.0-canary.73" : "0.0.0-test";
3396
+ var VERSION = true ? "4.0.0" : "0.0.0-test";
3397
3397
 
3398
3398
  // src/files/xai-files.ts
3399
3399
  import {