feat: add MiniMax as a chat completion provider (#5452) * feat: add MiniMax as a chat completion provider Add MiniMax (https://www.minimax.io) as a first-class chat completion provider. MiniMax already has TTS integration in SillyTavern; this extends support to LLM chat completions via their OpenAI-compatible API. Supported models: - MiniMax-M2.5 (default) — 204K context - MiniMax-M2.5-highspeed — same capability, faster inference Key implementation details: - Reuses existing SECRET_KEYS.MINIMAX (shared with TTS) - API endpoint: https://api.minimax.io/v1 - Temperature clamped to (0.0, 1.0] as required by MiniMax API - Returns hardcoded model list since MiniMax doesn't expose /v1/models - Full UI integration: model selector, sampler parameters, streaming Co-Authored-By: octo-patch <octo-patch@users.noreply.github.com> * feat: upgrade MiniMax default model to M2.7 - Add MiniMax-M2.7 and MiniMax-M2.7-highspeed to model list - Set MiniMax-M2.7 as default model - Keep all previous models as alternatives * feat: independent request function, vision support, temp clamping for MiniMax - Extract sendMinimaxRequest() following Chutes pattern (PR #4844) with function calling and JSON Schema structured output support - Clamp temperature to (0.01, 1.0] on backend; limit frontend UI max to 1.0 - Enable image inlining for MiniMax M2.7 model - Add MiniMax to slash-commands model selector and tokenizer mapping - Add minimax_model to default preset * feat: add VLM-based vision support for MiniMax M2.7 M2.7 does not natively accept image input. When images are detected in messages, pre-process them via the MiniMax VLM endpoint (/v1/coding_plan/vlm) to convert images to text descriptions before sending to the chat completions API. Uses the same API key. * feat: add M2-her model to MiniMax provider M2-her is MiniMax's dialogue/roleplay-optimized model with 64K context and 2048 max completion tokens. Text-only (no vision). * feat: add MiniMax China endpoint (minimaxi.com) support Add endpoint selector (Global/China) for MiniMax, mirroring the SiliconFlow pattern. Users can now choose between api.minimax.io (international) and api.minimaxi.com (China domestic). * fix: merge consecutive same-role messages for MiniMax MiniMax API rejects consecutive messages with the same role with error 'invalid chat setting (2013)'. Merge them before sending. * review: address PR feedback on MiniMax provider Backend (src/endpoints/backends/chat-completions.js): - Drop the entire MiniMax VLM image-preprocessing path; vision is no longer advertised for this provider, so M2.7 messages now go straight to /chat/completions without a separate VLM round-trip. - Drop the json_schema -> response_format mapping (MiniMax does not document structured-output support; relying on it was speculative). - Drop the backend temperature clamp; the same clamp now lives in the frontend so the wire payload matches what the user sees. - Drop the MINIMAX branch in /status that returned a hard-coded model list; the frontend hardcodes the same list and bypasses /status via noValidateSources, so the round-trip was wasted. - Add a streaming Transform + non-streaming helper that move <think>...</think> blocks from delta.content / message.content to reasoning_content. MiniMax M2.x emit chain-of-thought inline in content; without this transform the raw <think> tags leak into the rendered chat. Includes a state machine that holds back partial marker bytes so a marker split across SSE chunks is still detected. Frontend: - public/scripts/openai.js: add MINIMAX to noValidateSources so the key is accepted without a /models call; remove the dead saveModelList branch; clamp temperature to (0.0, 1.0] in createGenerationParameters. - public/scripts/reasoning.js: add MINIMAX to the non-streaming reasoning_content extraction case (the backend transform now produces this field for MiniMax responses). - public/scripts/slash-commands.js: add MINIMAX to the /api enum and add a MiniMax case to /api-url so users can switch endpoint by command. - public/scripts/custom-request.js: pass minimax_endpoint through the override-payload merge alongside the other per-source endpoint fields. - public/scripts/tokenizers.js: stop returning openai_model (which was always a MiniMax model id and thus an unknown tokenizer); fall back to gpt-3.5-turbo for a coarse but functional estimate. - public/scripts/tool-calling.js: add MINIMAX to supportedSources so function-calling settings are exposed. - public/index.html: drop the "-- Connect to the API --" placeholder option from the model select (the model list is hardcoded and always populated); remove minimax from the vision data-source attributes on the inline-media controls. - public/img/minimax.svg: replace the multicolor brand SVG with a single-color currentColor version that matches the other provider icons in the connect panel. * review: drop backend <think> parsing, defer to frontend Per reviewer feedback: SillyTavern's reasoningHandler / reasoning_auto_parse setting already extracts <think>...</think> blocks on the client side, so the backend doesn't need to rewrite MiniMax responses. Removes the SSE Transform, the non-streaming helper, and the corresponding case in reasoning.js. * fix: remove isImageInliningSupported declaration for MINIMAX * fix: remove MINIMAX from stream reasoning parsing * fix: add to autoconnect logic * fix: add missing MINIMAX models from docs * fix: freq. and pres. pen aren't supported for MINIMAX * fix: use clamp function for adjusting temperature * fix: pass minimax_endpoint from connection profile to ChatCompletionService * fix: update supported APIs in slash command documentation * fix: replace bespoke merge with standard MERGE_TOOLS processing * fix: add data-i18n attributes for headers --------- Co-authored-by: octo-patch <octo-patch@users.noreply.github.com> Co-authored-by: octo-patch <octo-patch@github.com> Co-authored-by: Cohee <18619528+Cohee1207@users.noreply.github.com>

aecbb9a2ee643ff6e49bee46a651f4b440ce67a5

Octopus <liyuan851277048@icloud.com>

Signed
Parents: a028bec87b53f617ff7aa3652ad0473bcb0d1529
13 files changed, +217 -8Ignore whitespace
default/content/presets/openai/Default.json+2 -0
@@ -10,6 +10,8 @@
1010 "mistralai_model": "mistral-large-latest",
1111 "chutes_model": "deepseek-ai/DeepSeek-V3-0324",
1212 "chutes_sort_models": "alphabetically",
13+ "minimax_model": "MiniMax-M2.7",
14+ "minimax_endpoint": "global",
1315 "electronhub_model": "gpt-4o-mini",
1416 "electronhub_sort_models": "alphabetically",
1517 "electronhub_group_models": false,
public/img/minimax.svg+1 -0
@@ -0,0 +1 @@
1+<svg fill="currentColor" fill-rule="evenodd" height="1em" style="flex:none;line-height:1" viewBox="0 0 24 24" width="1em" xmlns="http://www.w3.org/2000/svg"><title>Minimax</title><path d="M16.278 2c1.156 0 2.093.927 2.093 2.07v12.501a.74.74 0 00.744.709.74.74 0 00.743-.709V9.099a2.06 2.06 0 012.071-2.049A2.06 2.06 0 0124 9.1v6.561a.649.649 0 01-.652.645.649.649 0 01-.653-.645V9.1a.762.762 0 00-.766-.758.762.762 0 00-.766.758v7.472a2.037 2.037 0 01-2.048 2.026 2.037 2.037 0 01-2.048-2.026v-12.5a.785.785 0 00-.788-.753.785.785 0 00-.789.752l-.001 15.904A2.037 2.037 0 0113.441 22a2.037 2.037 0 01-2.048-2.026V18.04c0-.356.292-.645.652-.645.36 0 .652.289.652.645v1.934c0 .263.142.506.372.638.23.131.514.131.744 0a.734.734 0 00.372-.638V4.07c0-1.143.937-2.07 2.093-2.07zm-5.674 0c1.156 0 2.093.927 2.093 2.07v11.523a.648.648 0 01-.652.645.648.648 0 01-.652-.645V4.07a.785.785 0 00-.789-.78.785.785 0 00-.789.78v14.013a2.06 2.06 0 01-2.07 2.048 2.06 2.06 0 01-2.071-2.048V9.1a.762.762 0 00-.766-.758.762.762 0 00-.766.758v3.8a2.06 2.06 0 01-2.071 2.049A2.06 2.06 0 010 12.9v-1.378c0-.357.292-.646.652-.646.36 0 .653.29.653.646V12.9c0 .418.343.757.766.757s.766-.339.766-.757V9.099a2.06 2.06 0 012.07-2.048 2.06 2.06 0 012.071 2.048v8.984c0 .419.343.758.767.758.423 0 .766-.339.766-.758V4.07c0-1.143.937-2.07 2.093-2.07z"></path></svg>
1 \ No newline at end of file
public/index.html+30 -3
@@ -697,7 +697,7 @@
697697 </span>
698698 </div>
699699 </div>
700700 <div class="range-block" data-source="openai,claude,aimlapi,openrouter,ai21,makersuite,vertexai,mistralai,custom,cohere,perplexity,groq,siliconflow,minimax,electronhub,chutes,nanogpt,deepseek,xai,pollinations,moonshot,fireworks,cometapi,azure_openai,zai,workers_ai">
701701 <div class="range-block-title" data-i18n="Temperature">
702702 Temperature
703703 </div>
@@ -749,7 +749,7 @@
749749 </div>
750750 </div>
751751 </div>
752752 <div class="range-block" data-source="openai,claude,aimlapi,openrouter,ai21,makersuite,vertexai,mistralai,custom,cohere,perplexity,groq,siliconflow,minimax,electronhub,chutes,nanogpt,deepseek,xai,pollinations,moonshot,fireworks,cometapi,azure_openai,zai,workers_ai">
753753 <div class="range-block-title" data-i18n="Top P">
754754 Top P
755755 </div>
@@ -1997,7 +1997,7 @@
19971997 </b>
19981998 </div>
19991999 </div>
20002000 <div class="range-block" data-source="openai,cohere,mistralai,custom,claude,aimlapi,openrouter,groq,siliconflow,minimax,deepseek,makersuite,vertexai,ai21,xai,pollinations,moonshot,fireworks,cometapi,electronhub,chutes,azure_openai,zai,nanogpt,workers_ai">
20012001 <label for="openai_function_calling" class="checkbox_label flexWrap widthFreeExpand">
20022002 <input id="openai_function_calling" type="checkbox" />
20032003 <span data-i18n="Enable function calling">Enable function calling</span>
@@ -2907,6 +2907,7 @@
29072907 <option value="makersuite">Google AI Studio</option>
29082908 <option value="vertexai">Google Vertex AI</option>
29092909 <option value="mistralai">MistralAI</option>
2910+ <option value="minimax">MiniMax</option>
29102911 <option value="moonshot">Moonshot AI</option>
29112912 <option value="nanogpt">NanoGPT</option>
29122913 <option value="openrouter">OpenRouter</option>
@@ -3571,6 +3572,32 @@
35713572 <option value="" data-i18n="-- Connect to the API --">-- Connect to the API --</option>
35723573 </select>
35733574 </div>
3575+ <div id="minimax_form" data-source="minimax">
3576+ <h4 data-i18n="MiniMax API Key">MiniMax API Key</h4>
3577+ <div class="flex-container">
3578+ <input id="api_key_minimax" name="api_key_minimax" class="text_pole flex1" value="" type="text" autocomplete="off">
3579+ <div title="Manage API keys" data-i18n="[title]Manage API keys" class="menu_button fa-solid fa-key fa-fw manage-api-keys" data-key="api_key_minimax"></div>
3580+ </div>
3581+ <div data-for="api_key_minimax" class="neutral_warning" data-i18n="For privacy reasons, your API key will be hidden after you click 'Connect'.">
3582+ For privacy reasons, your API key will be hidden after you click 'Connect'.
3583+ </div>
3584+ <h4 data-i18n="MiniMax Endpoint">MiniMax Endpoint</h4>
3585+ <select id="minimax_endpoint">
3586+ <option value="global" data-i18n="Global (minimax.io)">Global (minimax.io)</option>
3587+ <option value="cn" data-i18n="China (minimaxi.com)">China (minimaxi.com)</option>
3588+ </select>
3589+ <h4 data-i18n="MiniMax Model">MiniMax Model</h4>
3590+ <select id="model_minimax_select">
3591+ <option value="MiniMax-M2.7">MiniMax-M2.7</option>
3592+ <option value="MiniMax-M2.7-highspeed">MiniMax-M2.7-highspeed</option>
3593+ <option value="MiniMax-M2.5">MiniMax-M2.5</option>
3594+ <option value="MiniMax-M2.5-highspeed">MiniMax-M2.5-highspeed</option>
3595+ <option value="MiniMax-M2.1">MiniMax-M2.1</option>
3596+ <option value="MiniMax-M2.1-highspeed">MiniMax-M2.1-highspeed</option>
3597+ <option value="MiniMax-M2">MiniMax-M2</option>
3598+ <option value="M2-her">M2-her</option>
3599+ </select>
3600+ </div>
35743601 <div id="electronhub_form" data-source="electronhub">
35753602 <h4 data-i18n="Electron Hub API Key">Electron Hub API Key</h4>
35763603 <div>
public/scripts/RossAscends-mods.js+1 -0
@@ -408,6 +408,7 @@ function RA_autoconnect(PrevApi) {
408408 || (secret_state[SECRET_KEYS.ZAI] && oai_settings.chat_completion_source == chat_completion_sources.ZAI)
409409 || (secret_state[SECRET_KEYS.POLLINATIONS] && oai_settings.chat_completion_source === chat_completion_sources.POLLINATIONS)
410410 || (secret_state[SECRET_KEYS.WORKERS_AI] && oai_settings.chat_completion_source == chat_completion_sources.WORKERS_AI)
411+ || (secret_state[SECRET_KEYS.MINIMAX] && oai_settings.chat_completion_source == chat_completion_sources.MINIMAX)
411412 || (isValidUrl(oai_settings.custom_url) && oai_settings.chat_completion_source == chat_completion_sources.CUSTOM)
412413 || (secret_state[SECRET_KEYS.AZURE_OPENAI] && oai_settings.chat_completion_source == chat_completion_sources.AZURE_OPENAI)
413414 ) {
public/scripts/custom-request.js+1 -1
@@ -591,7 +591,7 @@ export class ChatCompletionService {
591591 }
592592
593593 // Ensure api-url is properly applied for all sources that accept it
594594 ['custom_url', 'vertexai_region', 'zai_endpoint', 'siliconflow_endpoint', 'minimax_endpoint'].forEach(field => {
595595 // The order is: connection profile => CC preset => CC settings
596596 overridePayload[field] = overridePayload[field] || settings[field] || oai_settings[field];
597597 });
public/scripts/extensions/shared.js+1 -0
@@ -448,6 +448,7 @@ export class ConnectionManagerRequestService {
448448 vertexai_region: profile['api-url'],
449449 zai_endpoint: profile['api-url'],
450450 siliconflow_endpoint: profile['api-url'],
451+ minimax_endpoint: profile['api-url'],
451452 reverse_proxy: proxyPreset?.url,
452453 proxy_password: proxyPreset?.password,
453454 custom_prompt_post_processing: profile['prompt-post-processing'],
public/scripts/openai.js+52 -0
@@ -47,6 +47,7 @@ import { SECRET_KEYS, secret_state, writeSecret } from './secrets.js';
4747
4848import { getEventSourceStream } from './sse-stream.js';
4949import {
50+ clamp,
5051 createThumbnail,
5152 delay,
5253 download,
@@ -197,6 +198,7 @@ export const chat_completion_sources = {
197198 ZAI: 'zai',
198199 SILICONFLOW: 'siliconflow',
199200 WORKERS_AI: 'workers_ai',
201+ MINIMAX: 'minimax',
200202};
201203
202204const character_names_behavior = {
@@ -270,6 +272,11 @@ export const SILICONFLOW_ENDPOINT = {
270272 CN: 'cn',
271273};
272274
275+export const MINIMAX_ENDPOINT = {
276+ GLOBAL: 'global',
277+ CN: 'cn',
278+};
279+
273280const sensitiveFields = [
274281 'reverse_proxy',
275282 'proxy_password',
@@ -319,6 +326,8 @@ export const settingsToUpdate = {
319326 chutes_sort_models: ['#chutes_sort_models', 'chutes_sort_models', false, true],
320327 siliconflow_model: ['#model_siliconflow_select', 'siliconflow_model', false, true],
321328 siliconflow_endpoint: ['#siliconflow_endpoint', 'siliconflow_endpoint', false, true],
329+ minimax_model: ['#model_minimax_select', 'minimax_model', false, true],
330+ minimax_endpoint: ['#minimax_endpoint', 'minimax_endpoint', false, true],
322331 electronhub_model: ['#model_electronhub_select', 'electronhub_model', false, true],
323332 electronhub_sort_models: ['#electronhub_sort_models', 'electronhub_sort_models', false, true],
324333 electronhub_group_models: ['#electronhub_group_models', 'electronhub_group_models', false, true],
@@ -431,6 +440,8 @@ const default_settings = {
431440 chutes_sort_models: 'alphabetically',
432441 siliconflow_model: 'deepseek-ai/DeepSeek-V3',
433442 siliconflow_endpoint: SILICONFLOW_ENDPOINT.GLOBAL,
443+ minimax_model: 'MiniMax-M2.7',
444+ minimax_endpoint: MINIMAX_ENDPOINT.GLOBAL,
434445 electronhub_model: 'gpt-4o-mini',
435446 electronhub_sort_models: 'alphabetically',
436447 electronhub_group_models: false,
@@ -1712,6 +1723,8 @@ export function getChatCompletionModel(settings = null) {
17121723 return settings.groq_model;
17131724 case chat_completion_sources.SILICONFLOW:
17141725 return settings.siliconflow_model;
1726+ case chat_completion_sources.MINIMAX:
1727+ return settings.minimax_model;
17151728 case chat_completion_sources.ELECTRONHUB:
17161729 return settings.electronhub_model;
17171730 case chat_completion_sources.CHUTES:
@@ -2845,6 +2858,14 @@ export async function createGenerationParameters(settings, model, type, messages
28452858 generate_data.siliconflow_endpoint = settings.siliconflow_endpoint || SILICONFLOW_ENDPOINT.GLOBAL;
28462859 }
28472860
2861+ if (settings.chat_completion_source === chat_completion_sources.MINIMAX) {
2862+ generate_data.minimax_endpoint = settings.minimax_endpoint || MINIMAX_ENDPOINT.GLOBAL;
2863+ // MiniMax requires temperature in (0.0, 1.0]; zero is rejected.
2864+ if (Number.isFinite(generate_data.temperature)) {
2865+ generate_data.temperature = clamp(generate_data.temperature, Number.EPSILON, 1.0);
2866+ }
2867+ }
2868+
28482869 if (settings.chat_completion_source === chat_completion_sources.WORKERS_AI) {
28492870 generate_data.workers_ai_account_id = settings.workers_ai_account_id;
28502871 generate_data.top_k = settings.top_k_openai > 0 ? Math.min(Number(settings.top_k_openai), 50) : undefined;
@@ -4255,6 +4276,7 @@ async function getStatusOpen() {
42554276 chat_completion_sources.VERTEXAI,
42564277 chat_completion_sources.PERPLEXITY,
42574278 chat_completion_sources.ZAI,
4279+ chat_completion_sources.MINIMAX,
42584280 ];
42594281 if (noValidateSources.includes(oai_settings.chat_completion_source)) {
42604282 let status = t`Key saved; press \"Test Message\" to verify.`;
@@ -4312,6 +4334,10 @@ async function getStatusOpen() {
43124334 data.siliconflow_endpoint = oai_settings.siliconflow_endpoint;
43134335 }
43144336
4337+ if (oai_settings.chat_completion_source === chat_completion_sources.MINIMAX) {
4338+ data.minimax_endpoint = oai_settings.minimax_endpoint;
4339+ }
4340+
43154341 if (oai_settings.chat_completion_source === chat_completion_sources.WORKERS_AI) {
43164342 data.workers_ai_account_id = oai_settings.workers_ai_account_id;
43174343 }
@@ -5266,6 +5292,15 @@ async function onModelChange() {
52665292 oai_settings.siliconflow_model = value;
52675293 }
52685294
5295+ if ($(this).is('#model_minimax_select')) {
5296+ if (!value) {
5297+ console.debug('Null MiniMax model selected. Ignoring.');
5298+ return;
5299+ }
5300+ console.log('MiniMax model changed to', value);
5301+ oai_settings.minimax_model = value;
5302+ }
5303+
52695304 if ($(this).is('#model_electronhub_select')) {
52705305 if (!value || !hasModelsLoaded) {
52715306 console.debug('Null ElectronHub model selected. Ignoring.');
@@ -5707,6 +5742,15 @@ async function onModelChange() {
57075742 $('#temp_openai').attr('max', oai_max_temp).val(oai_settings.temp_openai).trigger('input');
57085743 }
57095744
5745+ if (oai_settings.chat_completion_source === chat_completion_sources.MINIMAX) {
5746+ const maxContext = oai_settings.minimax_model === 'M2-her' ? 65536 : 204800;
5747+ $('#openai_max_context').attr('max', maxContext);
5748+ oai_settings.openai_max_context = Math.min(Number($('#openai_max_context').attr('max')), oai_settings.openai_max_context);
5749+ $('#openai_max_context').val(oai_settings.openai_max_context).trigger('input');
5750+ oai_settings.temp_openai = Math.min(claude_max_temp, oai_settings.temp_openai);
5751+ $('#temp_openai').attr('max', claude_max_temp).val(oai_settings.temp_openai).trigger('input');
5752+ }
5753+
57105754 if (oai_settings.chat_completion_source == chat_completion_sources.ZAI) {
57115755 const maxContext = getZaiMaxContext(oai_settings.zai_model, oai_settings.max_context_unlocked);
57125756 $('#openai_max_context').attr('max', maxContext);
@@ -5780,6 +5824,7 @@ async function onConnectButtonClick(e) {
57805824 [chat_completion_sources.CHUTES]: { key: SECRET_KEYS.CHUTES, selector: '#api_key_chutes', proxy: false },
57815825 [chat_completion_sources.POLLINATIONS]: { key: SECRET_KEYS.POLLINATIONS, selector: '#api_key_pollinations', proxy: false },
57825826 [chat_completion_sources.WORKERS_AI]: { key: SECRET_KEYS.WORKERS_AI, selector: '#api_key_workers_ai', proxy: false },
5827+ [chat_completion_sources.MINIMAX]: { key: SECRET_KEYS.MINIMAX, selector: '#api_key_minimax', proxy: false },
57835828 };
57845829
57855830 // Vertex AI Express version - use API key
@@ -5845,6 +5890,8 @@ function toggleChatCompletionForms() {
58455890 $('#model_chutes_select').trigger('change');
58465891 } else if (oai_settings.chat_completion_source == chat_completion_sources.SILICONFLOW) {
58475892 $('#model_siliconflow_select').trigger('change');
5893+ } else if (oai_settings.chat_completion_source == chat_completion_sources.MINIMAX) {
5894+ $('#model_minimax_select').trigger('change');
58485895 } else if (oai_settings.chat_completion_source == chat_completion_sources.ELECTRONHUB) {
58495896 $('#model_electronhub_select').trigger('change');
58505897 } else if (oai_settings.chat_completion_source == chat_completion_sources.NANOGPT) {
@@ -7028,6 +7075,10 @@ export function initOpenAI() {
70287075 oai_settings.siliconflow_endpoint = String($(this).val());
70297076 saveSettingsDebounced();
70307077 });
7078+ $('#minimax_endpoint').on('input', function () {
7079+ oai_settings.minimax_endpoint = String($(this).val());
7080+ saveSettingsDebounced();
7081+ });
70317082 $('#workers_ai_account_id').on('input', function () {
70327083 oai_settings.workers_ai_account_id = String($(this).val());
70337084 saveSettingsDebounced();
@@ -7048,6 +7099,7 @@ export function initOpenAI() {
70487099 $('#model_groq_select').on('change', onModelChange);
70497100 $('#model_chutes_select').on('change', onModelChange);
70507101 $('#model_siliconflow_select').on('change', onModelChange);
7102+ $('#model_minimax_select').on('change', onModelChange);
70517103 $('#model_electronhub_select').on('change', onModelChange);
70527104 $('#model_nanogpt_select').on('change', onModelChange);
70537105 $('#model_deepseek_select').on('change', onModelChange);
public/scripts/secrets.js+2 -1
@@ -130,7 +130,7 @@ const FRIENDLY_NAMES = {
130130 [SECRET_KEYS.LINGVA_URL]: 'Lingva Endpoint (e.g. https://lingva.ml/api/v1)',
131131 [SECRET_KEYS.ONERING_URL]: 'OneRingTranslator Endpoint (e.g. http://127.0.0.1:4990/translate)',
132132 [SECRET_KEYS.DEEPLX_URL]: 'DeepLX Endpoint (e.g. http://127.0.0.1:1188/translate)',
133133 [SECRET_KEYS.MINIMAX]: 'MiniMax TTS',
134134 [SECRET_KEYS.MINIMAX_GROUP_ID]: 'MiniMax Group ID',
135135 [SECRET_KEYS.MOONSHOT]: 'Moonshot AI',
136136 [SECRET_KEYS.COMETAPI]: 'CometAPI',
@@ -184,6 +184,7 @@ const INPUT_MAP = {
184184 [SECRET_KEYS.AZURE_OPENAI]: '#api_key_azure_openai',
185185 [SECRET_KEYS.ZAI]: '#api_key_zai',
186186 [SECRET_KEYS.SILICONFLOW]: '#api_key_siliconflow',
187+ [SECRET_KEYS.MINIMAX]: '#api_key_minimax',
187188 [SECRET_KEYS.POLLINATIONS]: '#api_key_pollinations',
188189 [SECRET_KEYS.WORKERS_AI]: '#api_key_workers_ai',
189190};
public/scripts/slash-commands.js+30 -2
@@ -70,7 +70,7 @@ import { hideChatMessageRange } from './chats.js';
7070import { getContext, saveMetadataDebounced } from './extensions.js';
7171import { getRegexedString, regex_placement } from './extensions/regex/engine.js';
7272import { findGroupMemberId, groups, is_group_generating, openGroupById, regenerateGroup, resetSelectedGroup, saveGroupChat, selected_group, getGroupMembers } from './group-chats.js';
7373import { chat_completion_sources, MINIMAX_ENDPOINT, oai_settings, promptManager, SILICONFLOW_ENDPOINT, ZAI_ENDPOINT } from './openai.js';
7474import { user_avatar } from './personas.js';
7575import { addEphemeralStoppingString, chat_styles, context_presets, flushEphemeralStoppingStrings, playMessageSound, power_user } from './power-user.js';
7676import { SERVER_INPUTS, textgen_types, textgenerationwebui_settings } from './textgen-settings.js';
@@ -3134,6 +3134,7 @@ export function initDefaultSlashCommands() {
31343134 new SlashCommandEnumValue('zai', 'Z.AI', enumTypes.getBasedOnIndex(UNIQUE_APIS.findIndex(x => x === 'zai')), 'Z'),
31353135 new SlashCommandEnumValue('vertexai', 'Google Vertex AI', enumTypes.getBasedOnIndex(UNIQUE_APIS.findIndex(x => x === 'vertexai')), 'V'),
31363136 new SlashCommandEnumValue('siliconflow', 'SiliconFlow', enumTypes.getBasedOnIndex(UNIQUE_APIS.findIndex(x => x === 'siliconflow')), 'S'),
3137+ new SlashCommandEnumValue('minimax', 'MiniMax', enumTypes.getBasedOnIndex(UNIQUE_APIS.findIndex(x => x === 'minimax')), 'M'),
31373138 new SlashCommandEnumValue('kobold', 'KoboldAI Classic', enumTypes.getBasedOnIndex(UNIQUE_APIS.findIndex(x => x === 'kobold')), 'K'),
31383139 ...Object.values(textgen_types).filter(api => Object.keys(SERVER_INPUTS).includes(api)).map(api => new SlashCommandEnumValue(api, null, enumTypes.getBasedOnIndex(UNIQUE_APIS.findIndex(x => x === 'textgenerationwebui')), 'T')),
31393140 ],
@@ -3167,7 +3168,7 @@ export function initDefaultSlashCommands() {
31673168 ${t`If a manual API is provided to <b>set</b> the URL, make sure to set <code>connect=false</code>, as auto-connect only works for the currently selected API, or consider switching to it with <code>/api</code> first.`}
31683169 </div>
31693170 <div>
31703171 ${t`This slash command works for most of the Text Completion sources, KoboldAI Classic, and also Custom OpenAI compatible, Z.AI, SiliconFlow, MiniMax, and Google Vertex AI for the Chat Completion sources. If unsure which APIs are supported, check the auto-completion of the optional <code>api</code> argument of this command.`}
31713172 </div>
31723173 `,
31733174 }));
@@ -6261,6 +6262,7 @@ function getModelOptions(quiet) {
62616262 { id: 'model_groq_select', api: 'openai', type: chat_completion_sources.GROQ },
62626263 { id: 'model_chutes_select', api: 'openai', type: chat_completion_sources.CHUTES },
62636264 { id: 'model_siliconflow_select', api: 'openai', type: chat_completion_sources.SILICONFLOW },
6265+ { id: 'model_minimax_select', api: 'openai', type: chat_completion_sources.MINIMAX },
62646266 { id: 'model_electronhub_select', api: 'openai', type: chat_completion_sources.ELECTRONHUB },
62656267 { id: 'model_nanogpt_select', api: 'openai', type: chat_completion_sources.NANOGPT },
62666268 { id: 'model_deepseek_select', api: 'openai', type: chat_completion_sources.DEEPSEEK },
@@ -6628,6 +6630,32 @@ async function setApiUrlCallback({ api = null, connect = 'true', quiet = 'false'
66286630 return oai_settings.siliconflow_endpoint || SILICONFLOW_ENDPOINT.GLOBAL;
66296631 }
66306632
6633+ const isCurrentlyMinimax = main_api === 'openai' && oai_settings.chat_completion_source === chat_completion_sources.MINIMAX;
6634+ if (api === chat_completion_sources.MINIMAX || (!api && isCurrentlyMinimax)) {
6635+ if (!url) {
6636+ return oai_settings.minimax_endpoint || MINIMAX_ENDPOINT.GLOBAL;
6637+ }
6638+
6639+ const permittedValues = Object.values(MINIMAX_ENDPOINT);
6640+ if (!permittedValues.includes(url)) {
6641+ !isQuiet && toastr.warning(t`Valid options are: ${permittedValues.join(', ')}`, t`MiniMax endpoint '${url}' is not a valid option.`);
6642+ return '';
6643+ }
6644+
6645+ if (!isCurrentlyMinimax && autoConnect) {
6646+ toastr.warning(t`MiniMax is not the currently selected API, so we cannot do an auto-connect. Consider switching to it via /api beforehand.`);
6647+ return '';
6648+ }
6649+
6650+ $('#minimax_endpoint').val(url).trigger('input');
6651+
6652+ if (autoConnect) {
6653+ $('#api_button_openai').trigger('click');
6654+ }
6655+
6656+ return oai_settings.minimax_endpoint || MINIMAX_ENDPOINT.GLOBAL;
6657+ }
6658+
66316659 const isCurrentlyVertexAI = main_api === 'openai' && oai_settings.chat_completion_source === chat_completion_sources.VERTEXAI;
66326660 if (api === chat_completion_sources.VERTEXAI || (!api && isCurrentlyVertexAI)) {
66336661 const defaultRegion = 'us-central1';
public/scripts/tokenizers.js+5 -0
@@ -694,6 +694,11 @@ export function getTokenizerModel() {
694694 }
695695 }
696696
697+ if (oai_settings.chat_completion_source == chat_completion_sources.MINIMAX) {
698+ // MiniMax uses a proprietary tokenizer; fall back to a coarse OpenAI estimation.
699+ return 'gpt-3.5-turbo';
700+ }
701+
697702 if (oai_settings.chat_completion_source == chat_completion_sources.WORKERS_AI && oai_settings.workers_ai_model) {
698703 const model = oai_settings.workers_ai_model.toLowerCase();
699704
public/scripts/tool-calling.js+1 -0
@@ -667,6 +667,7 @@ export class ToolManager {
667667 chat_completion_sources.SILICONFLOW,
668668 chat_completion_sources.NANOGPT,
669669 chat_completion_sources.WORKERS_AI,
670+ chat_completion_sources.MINIMAX,
670671 ];
671672 return supportedSources.includes(settings.chat_completion_source);
672673 }
src/constants.js+6 -0
@@ -209,6 +209,7 @@ export const CHAT_COMPLETION_SOURCES = {
209209 AZURE_OPENAI: 'azure_openai',
210210 ZAI: 'zai',
211211 SILICONFLOW: 'siliconflow',
212+ MINIMAX: 'minimax',
212213 WORKERS_AI: 'workers_ai',
213214};
214215
@@ -557,3 +558,8 @@ export const SILICONFLOW_ENDPOINT = {
557558 GLOBAL: 'global',
558559 CN: 'cn',
559560};
561+
562+export const MINIMAX_ENDPOINT = {
563+ GLOBAL: 'global',
564+ CN: 'cn',
565+};
src/endpoints/backends/chat-completions.js+85 -1
@@ -18,6 +18,7 @@ import {
1818 OPENROUTER_HEADERS,
1919 VERTEX_SAFETY,
2020 SILICONFLOW_ENDPOINT,
21+ MINIMAX_ENDPOINT,
2122 ZAI_ENDPOINT,
2223} from '../../constants.js';
2324import {
@@ -89,6 +90,8 @@ const API_ZAI_COMMON = 'https://api.z.ai/api/paas/v4';
8990const API_ZAI_CODING = 'https://api.z.ai/api/coding/paas/v4';
9091const API_SILICONFLOW = 'https://api.siliconflow.com/v1';
9192const API_SILICONFLOW_CN = 'https://api.siliconflow.cn/v1';
93+const API_MINIMAX = 'https://api.minimax.io/v1';
94+const API_MINIMAX_CN = 'https://api.minimaxi.com/v1';
9295const API_OPENROUTER = 'https://openrouter.ai/api/v1';
9396const API_WORKERS_AI = 'https://api.cloudflare.com/client/v4/accounts';
9497
@@ -1552,7 +1555,87 @@ async function sendChutesRequest(request, response) {
15521555}
15531556
15541557/**
15551558 * Sends a chat completion request to Azure OpenAIMiniMax.
1559+ * @param {express.Request} request Express request
1560+ * @param {express.Response} response Express response
1561+ */
1562+async function sendMinimaxRequest(request, response) {
1563+ const apiUrl = request.body.minimax_endpoint === MINIMAX_ENDPOINT.CN
1564+ ? API_MINIMAX_CN : API_MINIMAX;
1565+ const apiKey = readSecret(request.user.directories, SECRET_KEYS.MINIMAX, request.body.secret_id);
1566+
1567+ if (!apiKey) {
1568+ console.warn('MiniMax key is missing.');
1569+ return response.status(400).send({ error: true });
1570+ }
1571+
1572+ const controller = new AbortController();
1573+ request.socket.removeAllListeners('close');
1574+ request.socket.on('close', function () {
1575+ controller.abort();
1576+ });
1577+
1578+ try {
1579+ // MiniMax does not allow consecutive messages with the same role.
1580+ // Merge them into a single message to avoid "invalid chat setting (2013)".
1581+ const messages = postProcessPrompt(request.body.messages, PROMPT_PROCESSING_TYPE.MERGE_TOOLS, getPromptNames(request));
1582+
1583+ let bodyParams = {};
1584+
1585+ if (Array.isArray(request.body.tools) && request.body.tools.length > 0) {
1586+ bodyParams['tools'] = request.body.tools;
1587+ bodyParams['tool_choice'] = request.body.tool_choice;
1588+ }
1589+
1590+ const requestBody = {
1591+ 'messages': messages,
1592+ 'model': request.body.model,
1593+ 'temperature': request.body.temperature,
1594+ 'max_tokens': request.body.model === 'M2-her' ? Math.min(request.body.max_tokens, 2048) : request.body.max_tokens,
1595+ 'stream': request.body.stream,
1596+ 'top_p': request.body.top_p,
1597+ 'stop': request.body.stop,
1598+ ...bodyParams,
1599+ };
1600+
1601+ const config = {
1602+ method: 'POST',
1603+ headers: {
1604+ 'Content-Type': 'application/json',
1605+ 'Authorization': 'Bearer ' + apiKey,
1606+ },
1607+ body: JSON.stringify(requestBody),
1608+ signal: controller.signal,
1609+ };
1610+
1611+ console.debug('MiniMax request:', requestBody);
1612+
1613+ const generateResponse = await fetch(apiUrl + '/chat/completions', config);
1614+
1615+ if (request.body.stream) {
1616+ await forwardFetchResponse(generateResponse, response);
1617+ } else {
1618+ if (!generateResponse.ok) {
1619+ const errorText = await generateResponse.text();
1620+ console.warn('MiniMax returned error: ', errorText);
1621+ const errorJson = tryParse(errorText) ?? { error: true };
1622+ return response.status(500).send(errorJson);
1623+ }
1624+ const generateResponseJson = await generateResponse.json();
1625+ console.debug('MiniMax response:', generateResponseJson);
1626+ return response.send(generateResponseJson);
1627+ }
1628+ } catch (error) {
1629+ console.error('Error communicating with MiniMax: ', error);
1630+ if (!response.headersSent) {
1631+ response.send({ error: true });
1632+ } else {
1633+ response.end();
1634+ }
1635+ }
1636+}
1637+
1638+/**
15561639 * @param {express.Request} request Express request object (contains request.body with all generate_data)
15571640 * @param {express.Response} response Express response object
15581641 */
@@ -2096,6 +2179,7 @@ router.post('/generate', async function (request, response) {
20962179 case CHAT_COMPLETION_SOURCES.AIMLAPI: return await sendAimlapiRequest(request, response);
20972180 case CHAT_COMPLETION_SOURCES.XAI: return await sendXaiRequest(request, response);
20982181 case CHAT_COMPLETION_SOURCES.CHUTES: return await sendChutesRequest(request, response);
2182+ case CHAT_COMPLETION_SOURCES.MINIMAX: return await sendMinimaxRequest(request, response);
20992183 case CHAT_COMPLETION_SOURCES.ELECTRONHUB: return await sendElectronHubRequest(request, response);
21002184 case CHAT_COMPLETION_SOURCES.AZURE_OPENAI: return await sendAzureOpenAIRequest(request, response);
21012185 }