@molecule/api-ai-minimax 1.1.3 → 1.2.0

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/LICENSE CHANGED
@@ -1,3 +1,4 @@
1
+
1
2
  Apache License
2
3
  Version 2.0, January 2004
3
4
  http://www.apache.org/licenses/
@@ -34,21 +35,33 @@
34
35
 
35
36
  "Work" shall mean the work of authorship, whether in Source or
36
37
  Object form, made available under the License, as indicated by a
37
- copyright notice that is included in or attached to the work.
38
+ copyright notice that is included in or attached to the work
39
+ (an example is provided in the Appendix below).
38
40
 
39
41
  "Derivative Works" shall mean any work, whether in Source or Object
40
42
  form, that is based on (or derived from) the Work and for which the
41
43
  editorial revisions, annotations, elaborations, or other modifications
42
- represent, as a whole, an original work of authorship.
43
-
44
- "Contribution" shall mean any work of authorship, including the
45
- original version of the Work and any modifications or additions
46
- to that Work, that is intentionally submitted to the Licensor for
47
- inclusion in the Work by the copyright owner or by an individual or
48
- Legal Entity authorized to submit on behalf of the copyright owner.
44
+ represent, as a whole, an original work of authorship. For the purposes
45
+ of this License, Derivative Works shall not include works that remain
46
+ separable from, or merely link (or bind by name) to the interfaces of,
47
+ the Work and Derivative Works thereof.
48
+
49
+ "Contribution" shall mean any work of authorship, including
50
+ the original version of the Work and any modifications or additions
51
+ to that Work or Derivative Works thereof, that is intentionally
52
+ submitted to Licensor for inclusion in the Work by the copyright owner
53
+ or by an individual or Legal Entity authorized to submit on behalf of
54
+ the copyright owner. For the purposes of this definition, "submitted"
55
+ means any form of electronic, verbal, or written communication sent
56
+ to the Licensor or its representatives, including but not limited to
57
+ communication on electronic mailing lists, source code control systems,
58
+ and issue tracking systems that are managed by, or on behalf of, the
59
+ Licensor for the purpose of discussing and improving the Work, but
60
+ excluding communication that is conspicuously marked or otherwise
61
+ designated in writing by the copyright owner as "Not a Contribution."
49
62
 
50
63
  "Contributor" shall mean Licensor and any individual or Legal Entity
51
- on behalf of whom a Contribution has been received by the Licensor and
64
+ on behalf of whom a Contribution has been received by Licensor and
52
65
  subsequently incorporated within the Work.
53
66
 
54
67
  2. Grant of Copyright License. Subject to the terms and conditions of
@@ -61,8 +74,18 @@
61
74
  3. Grant of Patent License. Subject to the terms and conditions of
62
75
  this License, each Contributor hereby grants to You a perpetual,
63
76
  worldwide, non-exclusive, no-charge, royalty-free, irrevocable
64
- patent license to make, have made, use, offer to sell, sell, import,
65
- and otherwise transfer the Work.
77
+ (except as stated in this section) patent license to make, have made,
78
+ use, offer to sell, sell, import, and otherwise transfer the Work,
79
+ where such license applies only to those patent claims licensable
80
+ by such Contributor that are necessarily infringed by their
81
+ Contribution(s) alone or by combination of their Contribution(s)
82
+ with the Work to which such Contribution(s) was submitted. If You
83
+ institute patent litigation against any entity (including a
84
+ cross-claim or counterclaim in a lawsuit) alleging that the Work
85
+ or a Contribution incorporated within the Work constitutes direct
86
+ or contributory patent infringement, then any patent licenses
87
+ granted to You under this License for that Work shall terminate
88
+ as of the date such litigation is filed.
66
89
 
67
90
  4. Redistribution. You may reproduce and distribute copies of the
68
91
  Work or Derivative Works thereof in any medium, with or without
@@ -84,23 +107,87 @@
84
107
  (d) If the Work includes a "NOTICE" text file as part of its
85
108
  distribution, then any Derivative Works that You distribute must
86
109
  include a readable copy of the attribution notices contained
87
- within such NOTICE file.
88
-
89
- 5. Submission of Contributions.
110
+ within such NOTICE file, excluding those notices that do not
111
+ pertain to any part of the Derivative Works, in at least one
112
+ of the following places: within a NOTICE text file distributed
113
+ as part of the Derivative Works; within the Source form or
114
+ documentation, if provided along with the Derivative Works; or,
115
+ within a display generated by the Derivative Works, if and
116
+ wherever such third-party notices normally appear. The contents
117
+ of the NOTICE file are for informational purposes only and
118
+ do not modify the License. You may add Your own attribution
119
+ notices within Derivative Works that You distribute, alongside
120
+ or as an addendum to the NOTICE text from the Work, provided
121
+ that such additional attribution notices cannot be construed
122
+ as modifying the License.
123
+
124
+ You may add Your own copyright statement to Your modifications and
125
+ may provide additional or different license terms and conditions
126
+ for use, reproduction, or distribution of Your modifications, or
127
+ for any such Derivative Works as a whole, provided Your use,
128
+ reproduction, and distribution of the Work otherwise complies with
129
+ the conditions stated in this License.
130
+
131
+ 5. Submission of Contributions. Unless You explicitly state otherwise,
132
+ any Contribution intentionally submitted for inclusion in the Work
133
+ by You to the Licensor shall be under the terms and conditions of
134
+ this License, without any additional terms or conditions.
135
+ Notwithstanding the above, nothing herein shall supersede or modify
136
+ the terms of any separate license agreement you may have executed
137
+ with Licensor regarding such Contributions.
90
138
 
91
139
  6. Trademarks. This License does not grant permission to use the trade
92
- names, trademarks, service marks, or product names of the Licensor.
140
+ names, trademarks, service marks, or product names of the Licensor,
141
+ except as required for reasonable and customary use in describing the
142
+ origin of the Work and reproducing the content of the NOTICE file.
93
143
 
94
144
  7. Disclaimer of Warranty. Unless required by applicable law or
95
- agreed to in writing, Licensor provides the Work on an "AS IS" BASIS,
96
- WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND.
97
-
98
- 8. Limitation of Liability. In no event and under no legal theory shall
99
- any Contributor be liable to You for damages.
100
-
101
- 9. Accepting Warranty or Additional Liability.
102
-
103
- Copyright 2026 Molecule Dev, Inc.
145
+ agreed to in writing, Licensor provides the Work (and each
146
+ Contributor provides its Contributions) on an "AS IS" BASIS,
147
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
148
+ implied, including, without limitation, any warranties or conditions
149
+ of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
150
+ PARTICULAR PURPOSE. You are solely responsible for determining the
151
+ appropriateness of using or redistributing the Work and assume any
152
+ risks associated with Your exercise of permissions under this License.
153
+
154
+ 8. Limitation of Liability. In no event and under no legal theory,
155
+ whether in tort (including negligence), contract, or otherwise,
156
+ unless required by applicable law (such as deliberate and grossly
157
+ negligent acts) or agreed to in writing, shall any Contributor be
158
+ liable to You for damages, including any direct, indirect, special,
159
+ incidental, or consequential damages of any character arising as a
160
+ result of this License or out of the use or inability to use the
161
+ Work (including but not limited to damages for loss of goodwill,
162
+ work stoppage, computer failure or malfunction, or any and all
163
+ other commercial damages or losses), even if such Contributor
164
+ has been advised of the possibility of such damages.
165
+
166
+ 9. Accepting Warranty or Additional Liability. While redistributing
167
+ the Work or Derivative Works thereof, You may choose to offer,
168
+ and charge a fee for, acceptance of support, warranty, indemnity,
169
+ or other liability obligations and/or rights consistent with this
170
+ License. However, in accepting such obligations, You may act only
171
+ on Your own behalf and on Your sole responsibility, not on behalf
172
+ of any other Contributor, and only if You agree to indemnify,
173
+ defend, and hold each Contributor harmless for any liability
174
+ incurred by, or claims asserted against, such Contributor by reason
175
+ of your accepting any such warranty or additional liability.
176
+
177
+ END OF TERMS AND CONDITIONS
178
+
179
+ APPENDIX: How to apply the Apache License to your work.
180
+
181
+ To apply the Apache License to your work, attach the following
182
+ boilerplate notice, with the fields enclosed by brackets "[]"
183
+ replaced with your own identifying information. (Don't include
184
+ the brackets!) The text should be enclosed in the appropriate
185
+ comment syntax for the file format. We also recommend that a
186
+ file or class name and description of purpose be included on the
187
+ same "printed page" as the copyright notice for easier
188
+ identification within third-party archives.
189
+
190
+ Copyright [yyyy] [name of copyright owner]
104
191
 
105
192
  Licensed under the Apache License, Version 2.0 (the "License");
106
193
  you may not use this file except in compliance with the License.
package/README.md CHANGED
@@ -3,17 +3,17 @@ AUTO-GENERATED — DO NOT EDIT THIS FILE.
3
3
  Generated by `mlcl sync-docs` from the package's src/index.ts JSDoc + mlcl/registry.json.
4
4
  Edits here are overwritten on the next commit (molecule's pre-commit hook regenerates).
5
5
  To change this document, edit the module-level JSDoc in src/index.ts.
6
- Generated: 2026-10-01T18:20:47.843Z
6
+ Generated: 2026-10-16T23:28:28.746Z
7
7
  -->
8
8
 
9
9
  # @molecule/api-ai-minimax
10
10
 
11
+ MiniMax AI provider for molecule.dev.
12
+
11
13
  > **Auto-generated, AI-first package reference** for the [molecule.dev](https://molecule.dev) ecosystem.
12
14
  > It is written to be read by coding agents as much as by people, and is generated from this
13
15
  > package's source — edit `src/index.ts` JSDoc, not this file.
14
16
 
15
- MiniMax AI provider for molecule.dev.
16
-
17
17
  ## Quick Start
18
18
 
19
19
  ```typescript
@@ -113,20 +113,20 @@ function createProvider(config?: MiniMaxConfig): AIProvider
113
113
 
114
114
  ### Constants
115
115
 
116
- #### `aiMinimaxSecretDefinitions`
116
+ #### `provider`
117
117
 
118
- Secret definitions required by the MiniMax AI bond.
118
+ The provider implementation.
119
119
 
120
120
  ```typescript
121
- const aiMinimaxSecretDefinitions: SecretDefinition[]
121
+ const provider: AIProvider
122
122
  ```
123
123
 
124
- #### `provider`
124
+ #### `aiMinimaxSecretDefinitions`
125
125
 
126
- The provider implementation.
126
+ Secret definitions required by the MiniMax AI bond.
127
127
 
128
128
  ```typescript
129
- const provider: AIProvider
129
+ const aiMinimaxSecretDefinitions: SecretDefinition[]
130
130
  ```
131
131
 
132
132
  ## Core Interface
@@ -1 +1 @@
1
- {"version":3,"file":"provider.d.ts","sourceRoot":"","sources":["../src/provider.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EACV,UAAU,EAOX,MAAM,kBAAkB,CAAA;AASzB,OAAO,cAAc,CAAA;AAErB,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AAmpB/C;;;;;GAKG;AACH,wBAAgB,cAAc,CAAC,MAAM,CAAC,EAAE,aAAa,GAAG,UAAU,CAEjE"}
1
+ {"version":3,"file":"provider.d.ts","sourceRoot":"","sources":["../src/provider.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAIH,OAAO,KAAK,EACV,UAAU,EAQX,MAAM,kBAAkB,CAAA;AASzB,OAAO,cAAc,CAAA;AAErB,OAAO,KAAK,EAAE,aAAa,EAAE,MAAM,YAAY,CAAA;AAmvB/C;;;;;GAKG;AACH,wBAAgB,cAAc,CAAC,MAAM,CAAC,EAAE,aAAa,GAAG,UAAU,CAEjE"}
package/dist/provider.js CHANGED
@@ -5,6 +5,7 @@
5
5
  *
6
6
  * @module
7
7
  */
8
+ import { randomUUID } from 'node:crypto';
8
9
  import { getLogger } from '@molecule/api-bond';
9
10
  import { t } from '@molecule/api-i18n';
10
11
  const logger = getLogger();
@@ -16,6 +17,8 @@ import './secrets.js';
16
17
  * Whether a model id is a MiniMax M3-family model. Only M3 accepts the
17
18
  * `thinking: {type: "adaptive" | "disabled"}` control — M2.x thinking is
18
19
  * always on and cannot be configured, and sending the param there is invalid.
20
+ * M3.1-Flash is also M3-family for multimodal gating, but see
21
+ * {@link isM31FlashFamily}: it REJECTS `disabled`.
19
22
  *
20
23
  * @param model - The model id (catalog `minimax-m3` or native `MiniMax-M3`).
21
24
  * @returns `true` for M3-family model ids.
@@ -23,6 +26,24 @@ import './secrets.js';
23
26
  function isM3Family(model) {
24
27
  return /^minimax-m3/i.test(model);
25
28
  }
29
+ /**
30
+ * Whether a model id is a MiniMax M3.1-Flash-family model. M3.1-Flash ALWAYS
31
+ * thinks: `thinking: {type: "disabled"}` is answered with a 400 ("requires
32
+ * adaptive thinking") and `reasoning_effort: "none"` is rejected too. Depth is
33
+ * tuned with `reasoning_effort` (`low|medium|high|xhigh|max`, API default
34
+ * `max`) — the only model family on this endpoint where that field does
35
+ * anything (others ignore it).
36
+ *
37
+ * Verified 2026-10-19 on platform.minimax.io (text-chat-openai reference +
38
+ * text-generation guide).
39
+ *
40
+ * @param model - The model id (catalog `minimax-m3.1-flash` or native
41
+ * `MiniMax-M3.1-Flash`).
42
+ * @returns `true` for M3.1-Flash-family model ids.
43
+ */
44
+ function isM31FlashFamily(model) {
45
+ return /^minimax-m3\.1/i.test(model);
46
+ }
26
47
  /**
27
48
  * MiniMax AI provider that implements the `AIProvider` interface
28
49
  * using MiniMax's OpenAI-compatible Chat Completions API with support
@@ -106,7 +127,7 @@ class MiniMaxAIProvider {
106
127
  body.tool_choice = { type: 'function', function: { name: params.toolChoice.name } };
107
128
  }
108
129
  }
109
- // MiniMax's native API has no reasoning_effort / thinking budget. M3 takes
130
+ // MiniMax's native API has no thinking budget. M3 takes
110
131
  // `thinking: {type: "adaptive" | "disabled"}` (the OpenAI-compatible
111
132
  // endpoint defaults to adaptive when omitted); M2.x thinking is always on
112
133
  // and takes no param. Synthase runs M3 as a non-thinking executor (like
@@ -115,13 +136,25 @@ class MiniMaxAIProvider {
115
136
  // re-hosts (`MiniMaxAI/MiniMax-M3`), which the `minimax-m3` anchor never
116
137
  // matches — that omission left thinking defaulted ON for every US turn.
117
138
  if (isM3Family(canonicalModel)) {
118
- body.thinking = { type: params.thinking ? 'adaptive' : 'disabled' };
119
- if (!params.thinking) {
120
- // Open-weights re-hosts (DeepInfra) IGNORE the native `thinking` param
121
- // and default reasoning ON — only `chat_template_kwargs.enable_thinking`
122
- // reaches the template there. The native host accepts both together
123
- // (verified live 2026-08-28), so send both and each host honors its own.
124
- body.chat_template_kwargs = { enable_thinking: false };
139
+ if (isM31FlashFamily(canonicalModel)) {
140
+ // M3.1-Flash ALWAYS thinks — `disabled` is a 400, so never send it
141
+ // (not even for tool-less calls that omit the thinking block). Depth
142
+ // via reasoning_effort when the caller resolved a level; omitted, the
143
+ // upstream default is `max`.
144
+ body.thinking = { type: 'adaptive' };
145
+ if (params.thinking?.effort) {
146
+ body.reasoning_effort = params.thinking.effort;
147
+ }
148
+ }
149
+ else {
150
+ body.thinking = { type: params.thinking ? 'adaptive' : 'disabled' };
151
+ if (!params.thinking) {
152
+ // Open-weights re-hosts (DeepInfra) IGNORE the native `thinking` param
153
+ // and default reasoning ON — only `chat_template_kwargs.enable_thinking`
154
+ // reaches the template there. The native host accepts both together
155
+ // (verified live 2026-08-28), so send both and each host honors its own.
156
+ body.chat_template_kwargs = { enable_thinking: false };
157
+ }
125
158
  }
126
159
  }
127
160
  // Keep thinking out of the visible message: with reasoning_split, thinking
@@ -131,9 +164,17 @@ class MiniMaxAIProvider {
131
164
  if (!params.thinking && params.temperature !== undefined) {
132
165
  body.temperature = params.temperature;
133
166
  }
167
+ // Per-call correlation (see ChatParams.requestId): a header rather than the
168
+ // body's `user`, which is the stable per-account abuse-attribution value.
169
+ const requestId = params.requestId ?? randomUUID();
170
+ const call = {
171
+ meta: { requestId, endpoint: `${this.baseUrl}${this.completionsPath}` },
172
+ endUserId: params.endUserId,
173
+ };
134
174
  const headers = {
135
175
  'Content-Type': 'application/json',
136
176
  Authorization: `Bearer ${this.apiKey}`,
177
+ 'X-Mol-Request-Id': requestId,
137
178
  };
138
179
  const DEFAULT_TIMEOUT_MS = 5 * 60 * 1000;
139
180
  const signal = params.signal ?? AbortSignal.timeout(DEFAULT_TIMEOUT_MS);
@@ -226,12 +267,20 @@ class MiniMaxAIProvider {
226
267
  yield { type: 'error', message: clientMessage, errorKey: 'ai.error.apiError' };
227
268
  return;
228
269
  }
270
+ // Whatever the transport exposes about this call — the handle a vendor
271
+ // investigation needs (DeepSeek's edge answers `x-ds-trace-id`).
272
+ const traceId = response.headers.get('x-request-id') ??
273
+ response.headers.get('request-id') ??
274
+ response.headers.get('x-ds-trace-id') ??
275
+ undefined;
276
+ if (traceId)
277
+ call.meta.traceId = traceId;
229
278
  if (params.stream === false) {
230
279
  const data = (await response.json());
231
280
  yield* this.parseNonStreamingResponse(data);
232
281
  return;
233
282
  }
234
- yield* this.parseStreamingResponse(response);
283
+ yield* this.parseStreamingResponse(response, call);
235
284
  }
236
285
  /**
237
286
  * Converts internal `ChatMessage` objects to the OpenAI-compatible message format.
@@ -410,7 +459,7 @@ class MiniMaxAIProvider {
410
459
  * @param response - The fetch Response with an SSE body stream.
411
460
  * @yields {ChatEvent} Chat events as they arrive from the SSE stream.
412
461
  */
413
- async *parseStreamingResponse(response) {
462
+ async *parseStreamingResponse(response, call) {
414
463
  const reader = response.body?.getReader();
415
464
  if (!reader) {
416
465
  yield {
@@ -434,7 +483,7 @@ class MiniMaxAIProvider {
434
483
  const lines = buffer.split('\n');
435
484
  buffer = lines.pop() ?? '';
436
485
  let yieldedContent = false;
437
- for (const event of this.processSSELines(lines, pendingTools, state)) {
486
+ for (const event of this.processSSELines(lines, pendingTools, state, call)) {
438
487
  yieldedContent = true;
439
488
  yield event;
440
489
  // A mid-stream error (rate-limit/overload during high load, sent as a
@@ -454,7 +503,7 @@ class MiniMaxAIProvider {
454
503
  }
455
504
  // Flush any remaining data in the buffer after EOF
456
505
  if (buffer.trim()) {
457
- for (const event of this.processSSELines(buffer.split('\n'), pendingTools, state)) {
506
+ for (const event of this.processSSELines(buffer.split('\n'), pendingTools, state, call)) {
458
507
  yield event;
459
508
  if (event.type === 'error')
460
509
  return;
@@ -487,6 +536,7 @@ class MiniMaxAIProvider {
487
536
  outputTokens: state.outputTokens,
488
537
  ...(state.cacheReadTokens ? { cacheReadInputTokens: state.cacheReadTokens } : {}),
489
538
  },
539
+ provider: call.meta,
490
540
  };
491
541
  }
492
542
  /**
@@ -496,9 +546,11 @@ class MiniMaxAIProvider {
496
546
  * @param lines - Raw SSE lines to process.
497
547
  * @param pendingTools - Mutable map tracking tool call deltas accumulated across chunks.
498
548
  * @param state - Mutable token usage counters.
549
+ * @param call - The request's identity: the answer's own `id` and `model` are
550
+ * recorded on `call.meta`, and any echo of the request is verified.
499
551
  * @yields {ChatEvent} Chat events extracted from the SSE lines.
500
552
  */
501
- *processSSELines(lines, pendingTools, state) {
553
+ *processSSELines(lines, pendingTools, state, call) {
502
554
  for (const line of lines) {
503
555
  if (!line.startsWith('data: '))
504
556
  continue;
@@ -507,6 +559,34 @@ class MiniMaxAIProvider {
507
559
  continue;
508
560
  try {
509
561
  const event = JSON.parse(json);
562
+ // The reply's own identity, recorded for the caller (see
563
+ // ProviderResponseMeta), and the request's identity verified against
564
+ // whatever the host echoes back. A reply carrying somebody else's
565
+ // identity is not ours to stream on — 2026-10-16 a conversation was
566
+ // handed a complete foreign completion and nothing could tell.
567
+ if (typeof event.id === 'string')
568
+ call.meta.responseId = event.id;
569
+ if (typeof event.model === 'string')
570
+ call.meta.model = event.model;
571
+ const echoedUser = typeof event.user === 'string' ? event.user : undefined;
572
+ const echoedRequest = typeof event.request_id === 'string'
573
+ ? event.request_id
574
+ : typeof event.requestId === 'string'
575
+ ? event.requestId
576
+ : undefined;
577
+ const wrongUser = echoedUser !== undefined && call.endUserId !== undefined && echoedUser !== call.endUserId;
578
+ const wrongRequest = echoedRequest !== undefined && echoedRequest !== call.meta.requestId;
579
+ if (wrongUser || wrongRequest) {
580
+ logger.error('Provider echoed a different identity — this reply is not ours', {
581
+ requestId: call.meta.requestId,
582
+ echoedUser: wrongUser ? echoedUser : undefined,
583
+ echoedRequestId: wrongRequest ? echoedRequest : undefined,
584
+ responseId: call.meta.responseId,
585
+ traceId: call.meta.traceId,
586
+ });
587
+ yield { type: 'error', message: t('ai.error.apiError'), errorKey: 'ai.error.apiError' };
588
+ return;
589
+ }
510
590
  // A mid-stream error arrives as a `data: {"error": {...}}` chunk (no
511
591
  // `choices`) — most importantly rate-limit/overload during high load.
512
592
  // Without this it matches no branch below and is silently dropped, so the
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@molecule/api-ai-minimax",
3
- "version": "1.1.3",
3
+ "version": "1.2.0",
4
4
  "description": "MiniMax AI provider for molecule.dev",
5
5
  "homepage": "https://www.molecule.dev/packages/api-ai-minimax",
6
6
  "type": "module",
@@ -33,7 +33,7 @@
33
33
  "directory": "packages/api/bonds/ai/minimax"
34
34
  },
35
35
  "devDependencies": {
36
- "@molecule/api-ai": "1.3.0",
36
+ "@molecule/api-ai": "1.3.4",
37
37
  "@molecule/api-bond": "1.0.2",
38
38
  "@types/node": "26.1.2",
39
39
  "typescript": "6.0.3",