@zerogpu/n8n-nodes-zerogpu 0.1.0 → 0.2.0

package/README.md

@@ -2,14 +2,13 @@
 
 Official [ZeroGPU](https://zerogpu.ai) community node package for [n8n](https://n8n.io/).
 
-> **Status: early release.** This package sets up the ZeroGPU integration for n8n, but it has no ZeroGPU operations yet. The ZeroGPU node shows a notice and passes input items through unchanged, and it doesn't call the ZeroGPU API.
-
 [n8n](https://n8n.io/) is a [fair-code licensed](https://docs.n8n.io/sustainable-use-license/) workflow automation platform.
 
 [About ZeroGPU](#about-zerogpu)  
 [Installation](#installation)  
 [Operations](#operations)  
 [Credentials](#credentials)  
+[Usage](#usage)  
 [Compatibility](#compatibility)  
 [Resources](#resources)  
 [License](#license)
@@ -34,11 +33,11 @@ Follow the [installation guide](https://docs.n8n.io/integrations/community-nodes
 
 ## Operations
 
-None yet. Model and inference operations will arrive in later releases.
+**ZeroGPU Chat Model** connects ZeroGPU's language models to n8n's AI nodes, such as AI Agent and Basic LLM Chain. More ZeroGPU models and operations will arrive in later releases.
 
 ## Credentials
 
-The package includes a **ZeroGPU API** credential, ready for future nodes:
+The package includes a **ZeroGPU API** credential, which the ZeroGPU Chat Model node uses to call ZeroGPU:
 
 1. Log in to the [ZeroGPU dashboard](https://platform.zerogpu.ai/dashboard).
 2. Copy a key from the **API Keys** page.
@@ -46,6 +45,35 @@ The package includes a **ZeroGPU API** credential, ready for future nodes:
 
 n8n checks the key when you save the credential. The check doesn't run any inference.
 
+## Usage
+
+1. Add an **AI Agent** or **Basic LLM Chain** node to your workflow.
+2. Connect a **ZeroGPU Chat Model** node to its **Chat Model** input and select your **ZeroGPU API** credential.
+3. Choose a model. The list comes from ZeroGPU, so it stays current as models are added. To use a model that isn't listed, switch **Model** to an expression and enter its ID. Model IDs are case-sensitive.
+
+Each setting under **Options** is sent only if you add it:
+
+| Option               | What it does                                       |
+| -------------------- | -------------------------------------------------- |
+| Maximum Tokens       | Limits the length of the response.                 |
+| Reasoning Effort     | Sets how much the model reasons before it answers. |
+| Sampling Temperature | Controls how random the response is.               |
+
+Some models have limits:
+
+- Models labelled **no tool calling** can't use AI Agent tools or return schema-validated JSON. Use them in a Basic LLM Chain, or choose another model for an agent with tools.
+- If you set **Maximum Tokens** or **Reasoning Effort** on a model that doesn't support it, the node stops with an error.
+- `gpt-6-luna` only accepts its default temperature, so leave **Sampling Temperature** unset.
+
+If n8n shows a general error, open its details to see ZeroGPU's message:
+
+| Message                                                                | What to do                                                      |
+| ---------------------------------------------------------------------- | --------------------------------------------------------------- |
+| `Invalid credentials`                                                  | Check the API key in your **ZeroGPU API** credential.           |
+| `Model not found`                                                      | Check the model ID, including its capitalization.               |
+| `Tool calling is not supported for model`                              | Choose a model that isn't labelled **no tool calling**.         |
+| `doesn't support Maximum Tokens` or `doesn't support Reasoning Effort` | Remove that setting under **Options**, or choose another model. |
+
 ## Compatibility
 
 Built against `n8n-workflow` 2.x. A minimum n8n version hasn't been established yet.

package/dist/nodes/ZeroGpuChatModel/ZeroGpuChatModel.node.d.ts

@@ -0,0 +1,10 @@
+import type { ILoadOptionsFunctions, INodePropertyOptions, INodeType, INodeTypeDescription, ISupplyDataFunctions, SupplyData } from 'n8n-workflow';
+export declare class ZeroGpuChatModel implements INodeType {
+    description: INodeTypeDescription;
+    methods: {
+        loadOptions: {
+            getModels(this: ILoadOptionsFunctions): Promise<INodePropertyOptions[]>;
+        };
+    };
+    supplyData(this: ISupplyDataFunctions, itemIndex: number): Promise<SupplyData>;
+}

package/dist/nodes/ZeroGpuChatModel/ZeroGpuChatModel.node.js

@@ -0,0 +1,165 @@
+"use strict";
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.ZeroGpuChatModel = void 0;
+const n8n_workflow_1 = require("n8n-workflow");
+const ai_node_sdk_1 = require("@n8n/ai-node-sdk");
+const BASE_URL = 'https://api.zerogpu.ai/v1';
+const MODELS_URL = 'https://api-dashboard.zerogpu.ai/api/models';
+async function fetchModels(context) {
+    var _a;
+    const response = (await context.helpers.httpRequest({
+        method: 'GET',
+        url: MODELS_URL,
+        json: true,
+    }));
+    return (_a = response.models) !== null && _a !== void 0 ? _a : [];
+}
+async function assertSettingsSupported(context, model, options, itemIndex) {
+    var _a;
+    let models;
+    try {
+        models = await fetchModels(context);
+    }
+    catch {
+        return;
+    }
+    const capabilities = (_a = models.find(({ modelId }) => modelId === model)) === null || _a === void 0 ? void 0 : _a.chat;
+    if (!capabilities)
+        return;
+    const unsupported = [];
+    if (options.maxTokens !== undefined && capabilities.maxTokens === false) {
+        unsupported.push('Maximum Tokens');
+    }
+    if (options.reasoningEffort !== undefined && capabilities.reasoningEffort === false) {
+        unsupported.push('Reasoning Effort');
+    }
+    if (unsupported.length > 0) {
+        throw new n8n_workflow_1.NodeOperationError(context.getNode(), `${model} doesn't support ${unsupported.join(' or ')}. Remove it from Options or choose another model.`, { itemIndex });
+    }
+}
+class ZeroGpuChatModel {
+    constructor() {
+        this.description = {
+            displayName: 'ZeroGPU Chat Model',
+            name: 'zeroGpuChatModel',
+            icon: { light: 'file:../../icons/zerogpu.svg', dark: 'file:../../icons/zerogpu.dark.svg' },
+            group: ['transform'],
+            version: [1],
+            subtitle: '={{$parameter["model"]}}',
+            description: 'Use ZeroGPU small and nano language models in AI workflows',
+            defaults: {
+                name: 'ZeroGPU Chat Model',
+            },
+            codex: {
+                categories: ['assistant'],
+                subcategories: {
+                    AI: ['Language Models', 'Root Nodes'],
+                    'Language Models': ['Chat Models (Recommended)'],
+                },
+                resources: {
+                    primaryDocumentation: [{ url: 'https://github.com/zerogpu/n8n-nodes-zerogpu#readme' }],
+                },
+            },
+            inputs: [],
+            outputs: [n8n_workflow_1.NodeConnectionTypes.AiLanguageModel],
+            outputNames: ['Model'],
+            credentials: [
+                {
+                    name: 'zeroGpuApi',
+                    required: true,
+                },
+            ],
+            properties: [
+                {
+                    displayName: 'Model Name or ID',
+                    name: 'model',
+                    type: 'options',
+                    required: true,
+                    typeOptions: {
+                        loadOptionsMethod: 'getModels',
+                        allowArbitraryValues: true,
+                    },
+                    default: 'deepseek-v4.1-flash',
+                    description: 'The model which will generate the completion. Choose from the list, or specify an ID using an <a href="https://docs.n8n.io/code/expressions/">expression</a>.',
+                },
+                {
+                    displayName: 'Options',
+                    name: 'options',
+                    placeholder: 'Add Option',
+                    description: 'Additional options to add',
+                    type: 'collection',
+                    default: {},
+                    options: [
+                        {
+                            displayName: 'Maximum Tokens',
+                            name: 'maxTokens',
+                            type: 'number',
+                            default: 1024,
+                            typeOptions: { minValue: 1 },
+                            description: 'The maximum number of tokens to generate in the response',
+                        },
+                        {
+                            displayName: 'Reasoning Effort',
+                            name: 'reasoningEffort',
+                            type: 'options',
+                            default: 'medium',
+                            options: [
+                                { name: 'Extra High', value: 'xhigh' },
+                                { name: 'High', value: 'high' },
+                                { name: 'Low', value: 'low' },
+                                { name: 'Medium', value: 'medium' },
+                                { name: 'Minimal', value: 'minimal' },
+                                { name: 'None', value: 'none' },
+                            ],
+                            description: 'How much the model reasons before it answers',
+                        },
+                        {
+                            displayName: 'Sampling Temperature',
+                            name: 'temperature',
+                            type: 'number',
+                            default: 0.7,
+                            typeOptions: { maxValue: 2, minValue: 0, numberPrecision: 1 },
+                            description: 'Controls randomness: Lowering results in less random completions. As the temperature approaches zero, the model will become deterministic and repetitive.',
+                        },
+                    ],
+                },
+            ],
+        };
+        this.methods = {
+            loadOptions: {
+                async getModels() {
+                    const models = await fetchModels(this);
+                    return models
+                        .filter(({ chat }) => chat)
+                        .map(({ modelId, modelDisplayName, pricing, chat }) => ({
+                        name: (chat === null || chat === void 0 ? void 0 : chat.tools) === false ? `${modelDisplayName} (no tool calling)` : modelDisplayName,
+                        value: modelId,
+                        description: pricing === null || pricing === void 0 ? void 0 : pricing.description,
+                    }))
+                        .sort((a, b) => a.name.localeCompare(b.name));
+                },
+            },
+        };
+    }
+    async supplyData(itemIndex) {
+        const model = this.getNodeParameter('model', itemIndex);
+        const options = this.getNodeParameter('options', itemIndex, {});
+        if (options.maxTokens !== undefined || options.reasoningEffort !== undefined) {
+            await assertSettingsSupported(this, model, options, itemIndex);
+        }
+        const credentials = await this.getCredentials('zeroGpuApi');
+        return (0, ai_node_sdk_1.supplyModel)(this, {
+            type: 'openai',
+            baseUrl: BASE_URL,
+            apiKey: credentials.apiKey,
+            model,
+            temperature: options.temperature,
+            maxTokens: options.maxTokens,
+            additionalParams: options.reasoningEffort
+                ? { reasoning_effort: options.reasoningEffort }
+                : undefined,
+        });
+    }
+}
+exports.ZeroGpuChatModel = ZeroGpuChatModel;
+//# sourceMappingURL=ZeroGpuChatModel.node.js.map

package/dist/nodes/ZeroGpuChatModel/ZeroGpuChatModel.node.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"ZeroGpuChatModel.node.js","sourceRoot":"","sources":["../../../nodes/ZeroGpuChatModel/ZeroGpuChatModel.node.ts"],"names":[],"mappings":";;;AAQA,+CAAuE;AACvE,kDAA+C;AAE/C,MAAM,QAAQ,GAAG,2BAA2B,CAAC;AAC7C,MAAM,UAAU,GAAG,6CAA6C,CAAC;AAgBjE,KAAK,UAAU,WAAW,CACzB,OAAqD;;IAErD,MAAM,QAAQ,GAAG,CAAC,MAAM,OAAO,CAAC,OAAO,CAAC,WAAW,CAAC;QACnD,MAAM,EAAE,KAAK;QACb,GAAG,EAAE,UAAU;QACf,IAAI,EAAE,IAAI;KACV,CAAC,CAAgC,CAAC;IAEnC,OAAO,MAAA,QAAQ,CAAC,MAAM,mCAAI,EAAE,CAAC;AAC9B,CAAC;AAGD,KAAK,UAAU,uBAAuB,CACrC,OAA6B,EAC7B,KAAa,EACb,OAAqB,EACrB,SAAiB;;IAEjB,IAAI,MAAsB,CAAC;IAC3B,IAAI,CAAC;QACJ,MAAM,GAAG,MAAM,WAAW,CAAC,OAAO,CAAC,CAAC;IACrC,CAAC;IAAC,MAAM,CAAC;QAER,OAAO;IACR,CAAC;IAED,MAAM,YAAY,GAAG,MAAA,MAAM,CAAC,IAAI,CAAC,CAAC,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC,OAAO,KAAK,KAAK,CAAC,0CAAE,IAAI,CAAC;IAC3E,IAAI,CAAC,YAAY;QAAE,OAAO;IAE1B,MAAM,WAAW,GAAa,EAAE,CAAC;IACjC,IAAI,OAAO,CAAC,SAAS,KAAK,SAAS,IAAI,YAAY,CAAC,SAAS,KAAK,KAAK,EAAE,CAAC;QACzE,WAAW,CAAC,IAAI,CAAC,gBAAgB,CAAC,CAAC;IACpC,CAAC;IACD,IAAI,OAAO,CAAC,eAAe,KAAK,SAAS,IAAI,YAAY,CAAC,eAAe,KAAK,KAAK,EAAE,CAAC;QACrF,WAAW,CAAC,IAAI,CAAC,kBAAkB,CAAC,CAAC;IACtC,CAAC;IAED,IAAI,WAAW,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,MAAM,IAAI,iCAAkB,CAC3B,OAAO,CAAC,OAAO,EAAE,EACjB,GAAG,KAAK,oBAAoB,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,mDAAmD,EACvG,EAAE,SAAS,EAAE,CACb,CAAC;IACH,CAAC;AACF,CAAC;AAED,MAAa,gBAAgB;IAA7B;QACC,gBAAW,GAAyB;YACnC,WAAW,EAAE,oBAAoB;YACjC,IAAI,EAAE,kBAAkB;YACxB,IAAI,EAAE,EAAE,KAAK,EAAE,8BAA8B,EAAE,IAAI,EAAE,mCAAmC,EAAE;YAC1F,KAAK,EAAE,CAAC,WAAW,CAAC;YACpB,OAAO,EAAE,CAAC,CAAC,CAAC;YACZ,QAAQ,EAAE,0BAA0B;YACpC,WAAW,EAAE,4DAA4D;YACzE,QAAQ,EAAE;gBACT,IAAI,EAAE,oBAAoB;aAC1B;YACD,KAAK,EAAE;gBACN,UAAU,EAAE,CAAC,WAAW,CAAC;gBACzB,aAAa,EAAE;oBACd,EAAE,EAAE,CAAC,iBAAiB,EAAE,YAAY,CAAC;oBACrC,iBAAiB,EAAE,CAAC,2BAA2B,CAAC;iBAChD;gBACD,SAAS,EAAE;oBACV,oBAAoB,EAAE,CAAC,EAAE,GAAG,EAAE,qDAAqD,EAAE,CAAC;iBACtF;aACD;YACD,MAAM,EAAE,EAAE;YACV,OAAO,EAAE,CAAC,kCAAmB,CAAC,eAAe,CAAC;YAC9C,WAAW,EAAE,CAAC,OAAO,CAAC;YACtB,WAAW,EAAE;gBACZ;oBACC,IAAI,EAAE,YAAY;oBAClB,QAAQ,EAAE,IAAI;iBACd;aACD;YACD,UAAU,EAAE;gBACX;oBACC,WAAW,EAAE,kBAAkB;oBAC/B,IAAI,EAAE,OAAO;oBACb,IAAI,EAAE,SAAS;oBACf,QAAQ,EAAE,IAAI;oBACd,WAAW,EAAE;wBACZ,iBAAiB,EAAE,WAAW;wBAC9B,oBAAoB,EAAE,IAAI;qBAC1B;oBACD,OAAO,EAAE,qBAAqB;oBAC9B,WAAW,EACV,+JAA+J;iBAChK;gBACD;oBACC,WAAW,EAAE,SAAS;oBACtB,IAAI,EAAE,SAAS;oBACf,WAAW,EAAE,YAAY;oBACzB,WAAW,EAAE,2BAA2B;oBACxC,IAAI,EAAE,YAAY;oBAClB,OAAO,EAAE,EAAE;oBACX,OAAO,EAAE;wBACR;4BACC,WAAW,EAAE,gBAAgB;4BAC7B,IAAI,EAAE,WAAW;4BACjB,IAAI,EAAE,QAAQ;4BACd,OAAO,EAAE,IAAI;4BACb,WAAW,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE;4BAC5B,WAAW,EAAE,0DAA0D;yBACvE;wBACD;4BACC,WAAW,EAAE,kBAAkB;4BAC/B,IAAI,EAAE,iBAAiB;4BACvB,IAAI,EAAE,SAAS;4BACf,OAAO,EAAE,QAAQ;4BACjB,OAAO,EAAE;gCACR,EAAE,IAAI,EAAE,YAAY,EAAE,KAAK,EAAE,OAAO,EAAE;gCACtC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE;gCAC/B,EAAE,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE;gCAC7B,EAAE,IAAI,EAAE,QAAQ,EAAE,KAAK,EAAE,QAAQ,EAAE;gCACnC,EAAE,IAAI,EAAE,SAAS,EAAE,KAAK,EAAE,SAAS,EAAE;gCACrC,EAAE,IAAI,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE;6BAC/B;4BACD,WAAW,EAAE,8CAA8C;yBAC3D;wBACD;4BACC,WAAW,EAAE,sBAAsB;4BACnC,IAAI,EAAE,aAAa;4BACnB,IAAI,EAAE,QAAQ;4BACd,OAAO,EAAE,GAAG;4BACZ,WAAW,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,eAAe,EAAE,CAAC,EAAE;4BAC7D,WAAW,EACV,2JAA2J;yBAC5J;qBACD;iBACD;aACD;SACD,CAAC;QAEF,YAAO,GAAG;YACT,WAAW,EAAE;gBACZ,KAAK,CAAC,SAAS;oBACd,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC,IAAI,CAAC,CAAC;oBAEvC,OAAO,MAAM;yBACX,MAAM,CAAC,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,IAAI,CAAC;yBAC1B,GAAG,CAAC,CAAC,EAAE,OAAO,EAAE,gBAAgB,EAAE,OAAO,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC;wBACvD,IAAI,EACH,CAAA,IAAI,aAAJ,IAAI,uBAAJ,IAAI,CAAE,KAAK,MAAK,KAAK,CAAC,CAAC,CAAC,GAAG,gBAAgB,oBAAoB,CAAC,CAAC,CAAC,gBAAgB;wBACnF,KAAK,EAAE,OAAO;wBACd,WAAW,EAAE,OAAO,aAAP,OAAO,uBAAP,OAAO,CAAE,WAAW;qBACjC,CAAC,CAAC;yBACF,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;gBAChD,CAAC;aACD;SACD,CAAC;IAyBH,CAAC;IAvBA,KAAK,CAAC,UAAU,CAA6B,SAAiB;QAC7D,MAAM,KAAK,GAAG,IAAI,CAAC,gBAAgB,CAAC,OAAO,EAAE,SAAS,CAAW,CAAC;QAClE,MAAM,OAAO,GAAG,IAAI,CAAC,gBAAgB,CAAC,SAAS,EAAE,SAAS,EAAE,EAAE,CAAiB,CAAC;QAEhF,IAAI,OAAO,CAAC,SAAS,KAAK,SAAS,IAAI,OAAO,CAAC,eAAe,KAAK,SAAS,EAAE,CAAC;YAC9E,MAAM,uBAAuB,CAAC,IAAI,EAAE,KAAK,EAAE,OAAO,EAAE,SAAS,CAAC,CAAC;QAChE,CAAC;QAED,MAAM,WAAW,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC,YAAY,CAAC,CAAC;QAE5D,OAAO,IAAA,yBAAW,EAAC,IAAI,EAAE;YACxB,IAAI,EAAE,QAAQ;YACd,OAAO,EAAE,QAAQ;YACjB,MAAM,EAAE,WAAW,CAAC,MAAgB;YACpC,KAAK;YACL,WAAW,EAAE,OAAO,CAAC,WAAW;YAChC,SAAS,EAAE,OAAO,CAAC,SAAS;YAE5B,gBAAgB,EAAE,OAAO,CAAC,eAAe;gBACxC,CAAC,CAAC,EAAE,gBAAgB,EAAE,OAAO,CAAC,eAAe,EAAE;gBAC/C,CAAC,CAAC,SAAS;SACZ,CAAC,CAAC;IACJ,CAAC;CACD;AAnID,4CAmIC"}

package/dist/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@zerogpu/n8n-nodes-zerogpu",
-    "version": "0.1.0",
+    "version": "0.2.0",
     "description": "Official ZeroGPU nodes for n8n. ZeroGPU is a compute-efficient inference provider for apps and agents — purpose-built small and nano language models on an edge network that run the repeatable tasks frontier models shouldn’t, ~10x faster and 50%+ cheaper. Auto-scaling, with zero GPU infrastructure. Plug in and you’re live.",
     "license": "MIT",
     "homepage": "https://zerogpu.ai",
@@ -39,11 +39,12 @@
     "n8n": {
         "n8nNodesApiVersion": 1,
         "strict": true,
+        "aiNodeSdkVersion": 1,
         "credentials": [
             "dist/credentials/ZeroGpuApi.credentials.js"
         ],
         "nodes": [
-            "dist/nodes/ZeroGpu/ZeroGpu.node.js"
+            "dist/nodes/ZeroGpuChatModel/ZeroGpuChatModel.node.js"
         ]
     },
     "devDependencies": {
@@ -54,6 +55,7 @@
         "typescript": "5.9.2"
     },
     "peerDependencies": {
-        "n8n-workflow": "*"
+        "n8n-workflow": "*",
+        "@n8n/ai-node-sdk": "*"
     }
 }