@coder-ka/ollama-as-webapi 1.0.0

package/README.md

@@ -0,0 +1,74 @@
+# ollama-as-webapi
+
+This repository provides a simple CLI tool to serve Ollama models as a Web API.
+
+## Installation
+
+```bash
+npm i -g @coder-ka/ollama-as-webapi
+```
+
+## Usage
+
+```bash
+ollama-as-webapi
+```
+
+### CLI Options
+
+| Option | Description | Default |
+| --- | --- | --- |
+| `--model` | Model name | `gemma3:4b` |
+| `--port` | Port number | `3000` |
+| `--host` | Host name | `localhost` |
+| `--gpu` | Use GPU | `true` |
+
+## Web API Reference
+
+### Content Generation
+
+`/generate`
+
+#### Request Body
+
+| Field | Required | Type | Description | Default |
+| --- | --- | --- | --- | --- |
+| `prompt` | Yes | string | Prompt to generate content for | `""` |
+| `maxTokens` | No | number | Maximum number of tokens to generate | `256` |
+| `temperature` | No | number | Temperature for sampling | `0.7` |
+| `topP` | No | number | Top-p sampling | `0.9` |
+| `topK` | No | number | Top-k sampling | `40` |
+| `n` | No | number | Number of samples to generate | `1` |
+| `stop` | No | string[] | Stop sequences | `[]` |
+
+#### Response Body
+
+| Field | Type | Description |
+| --- | --- | --- |
+| `text` | string | Generated content |
+| `tokens` | number | Number of tokens generated |
+| `finishReason` | string | Reason for finishing generation |
+
+#### Example
+
+```bash
+curl -X POST http://localhost:3000/generate \
+  -H "Content-Type: application/json" \
+  -d '{"prompt": "Hello"}'
+```
+
+### License
+
+MIT
+
+### How to contribute
+
+1. Fork the repository
+2. Create a feature branch
+3. Commit your changes
+4. Push to the branch
+5. Open a Pull Request
+
+### Author
+
+[Katsuyuki Oeda](https://github.com/coder-ka)

package/dist/cli.js

@@ -0,0 +1,26 @@
+import { Command } from 'commander';
+import { startServer } from './server.js';
+import { pullModelIfNeeded } from './ollama.js';
+const program = new Command();
+program
+    .name('ollama-as-webapi')
+    .description('Serve Ollama models as a Web API')
+    .version('1.0.0');
+program
+    .option('--model <name>', 'Model name', 'gemma3:4b')
+    .option('--port <number>', 'Port number', '3000')
+    .option('--host <name>', 'Host name', 'localhost')
+    .option('--gpu <boolean>', 'Use GPU', 'true')
+    .action(async (options) => {
+    const port = parseInt(options.port, 10);
+    const gpu = options.gpu === 'true';
+    try {
+        await pullModelIfNeeded(options.model);
+        startServer(options.host, port, options.model, gpu);
+    }
+    catch (error) {
+        console.error('Failed to start server:', error);
+        process.exit(1);
+    }
+});
+program.parse();

package/dist/ollama.js

@@ -0,0 +1,32 @@
+import ollama from 'ollama';
+export async function generateContent(options) {
+    // Note: Ollama handles GPU automatically based on system configuration and model.
+    // The 'gpu' option here is mostly for user preference if we were using a different backend,
+    // but with Ollama we'll pass model options.
+    const response = await ollama.generate({
+        model: options.model,
+        prompt: options.prompt,
+        options: {
+            num_predict: options.maxTokens,
+            temperature: options.temperature,
+            top_p: options.topP,
+            top_k: options.topK,
+            stop: options.stop,
+        },
+        stream: false,
+    });
+    return {
+        text: response.response,
+        tokens: response.eval_count || 0,
+        finishReason: 'stop', // Ollama response doesn't explicitly provide finish_reason in generate call in the same way OpenAI does, but 'stop' is standard for non-stream.
+    };
+}
+export async function pullModelIfNeeded(model) {
+    const models = await ollama.list();
+    const exists = models.models.some((m) => m.name === model || m.name.startsWith(model + ':'));
+    if (!exists) {
+        console.log(`Pulling model ${model}...`);
+        await ollama.pull({ model });
+        console.log(`Model ${model} pulled successfully.`);
+    }
+}

package/dist/server.js

@@ -0,0 +1,37 @@
+import { Hono } from 'hono';
+import { serve } from '@hono/node-server';
+import { generateContent } from './ollama.js';
+export function startServer(host, port, model, gpu) {
+    const app = new Hono();
+    app.post('/generate', async (c) => {
+        try {
+            const body = await c.req.json();
+            const options = {
+                model,
+                prompt: body.prompt || '',
+                maxTokens: body.maxTokens,
+                temperature: body.temperature,
+                topP: body.topP,
+                topK: body.topK,
+                stop: body.stop,
+                gpu,
+            };
+            if (!options.prompt) {
+                return c.json({ error: 'Prompt is required' }, 400);
+            }
+            const result = await generateContent(options);
+            return c.json(result);
+        }
+        catch (error) {
+            console.error('Error generating content:', error);
+            return c.json({ error: error.message || 'Internal Server Error' }, 500);
+        }
+    });
+    console.log(`Starting server on http://${host}:${port}`);
+    console.log(`Using model: ${model}`);
+    serve({
+        fetch: app.fetch,
+        port,
+        hostname: host,
+    });
+}

package/package.json

@@ -0,0 +1,48 @@
+{
+  "name": "@coder-ka/ollama-as-webapi",
+  "version": "1.0.0",
+  "description": "This repository provides a simple CLI tool to serve Ollama models as a Web API.",
+  "type": "module",
+  "bin": {
+    "ollama-as-webapi": "dist/cli.js"
+  },
+  "scripts": {
+    "build": "tsc",
+    "start": "node dist/cli.js",
+    "dev": "tsx src/cli.ts",
+    "prepublishOnly": "npm run build"
+  },
+  "author": "Katsuyuki Oeda",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/coder-ka/ollama-as-webapi.git"
+  },
+  "bugs": {
+    "url": "https://github.com/coder-ka/ollama-as-webapi/issues"
+  },
+  "homepage": "https://github.com/coder-ka/ollama-as-webapi#readme",
+  "keywords": [
+    "ollama",
+    "ai",
+    "llm",
+    "webapi",
+    "cli"
+  ],
+  "files": [
+    "dist",
+    "package.json"
+  ],
+  "dependencies": {
+    "@hono/node-server": "^1.19.9",
+    "commander": "^14.0.3",
+    "hono": "^4.12.0",
+    "ollama": "^0.6.3"
+  },
+  "devDependencies": {
+    "@types/commander": "^2.12.0",
+    "@types/node": "^25.3.0",
+    "tsx": "^4.21.0",
+    "typescript": "^5.9.3"
+  }
+}