Add Ollama provider for self-hosted LLM inference

Adds LLM_PROVIDER=ollama for fully local, zero-cost inference
via Ollama's OpenAI-compatible API. No API key required.
Configurable base URL via OLLAMA_BASE_URL env var.

test/llm-ollama-integration.test.mjs

@@ -0,0 +1,36 @@
+// Ollama provider — integration test (calls real Ollama instance)
+// Requires a running Ollama server with a model pulled
+// Run: OLLAMA_MODEL=llama3.1:8b node --test test/llm-ollama-integration.test.mjs
+
+import { describe, it } from 'node:test';
+import assert from 'node:assert/strict';
+import { OllamaProvider } from '../lib/llm/ollama.mjs';
+
+const BASE_URL = process.env.OLLAMA_BASE_URL || 'http://localhost:11434';
+const MODEL = process.env.OLLAMA_MODEL || 'llama3.1:8b';
+
+// Check if Ollama is reachable before running tests
+let ollamaAvailable = false;
+try {
+  const res = await fetch(`${BASE_URL}/api/tags`, { signal: AbortSignal.timeout(3000) });
+  ollamaAvailable = res.ok;
+} catch { /* not available */ }
+
+describe('Ollama integration', { skip: !ollamaAvailable && 'Ollama not reachable' }, () => {
+  it('should complete a prompt via local Ollama', async () => {
+    const provider = new OllamaProvider({ model: MODEL, baseUrl: BASE_URL });
+    assert.equal(provider.isConfigured, true);
+
+    const result = await provider.complete(
+      'You are a helpful assistant. Respond in exactly one sentence.',
+      'What is 2+2?',
+      { maxTokens: 128, timeout: 60000 }
+    );
+
+    assert.ok(result.text.length > 0, 'Response text should not be empty');
+    assert.ok(result.model, 'Should report model name');
+    console.log(`  Response: ${result.text}`);
+    console.log(`  Tokens: ${result.usage.inputTokens} in / ${result.usage.outputTokens} out`);
+    console.log(`  Model: ${result.model}`);
+  });
+});

test/llm-ollama.test.mjs

@@ -0,0 +1,170 @@
+// Ollama provider — unit tests
+// Uses Node.js built-in test runner (node:test) — no extra dependencies
+
+import { describe, it, mock, beforeEach } from 'node:test';
+import assert from 'node:assert/strict';
+import { OllamaProvider } from '../lib/llm/ollama.mjs';
+import { createLLMProvider } from '../lib/llm/index.mjs';
+
+// ─── Unit Tests ───
+
+describe('OllamaProvider', () => {
+  it('should set defaults correctly', () => {
+    const provider = new OllamaProvider({});
+    assert.equal(provider.name, 'ollama');
+    assert.equal(provider.model, 'llama3.1:8b');
+    assert.equal(provider.baseUrl, 'http://localhost:11434');
+    assert.equal(provider.isConfigured, true);
+  });
+
+  it('should accept custom model and base URL', () => {
+    const provider = new OllamaProvider({ model: 'qwen2.5:14b', baseUrl: 'http://192.168.1.10:11434' });
+    assert.equal(provider.model, 'qwen2.5:14b');
+    assert.equal(provider.baseUrl, 'http://192.168.1.10:11434');
+  });
+
+  it('should strip trailing slashes from base URL', () => {
+    const provider = new OllamaProvider({ baseUrl: 'http://localhost:11434/' });
+    assert.equal(provider.baseUrl, 'http://localhost:11434');
+  });
+
+  it('should throw on API error', async () => {
+    const provider = new OllamaProvider({});
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn(() =>
+      Promise.resolve({ ok: false, status: 404, text: () => Promise.resolve('model not found') })
+    );
+    try {
+      await assert.rejects(
+        () => provider.complete('system', 'user'),
+        (err) => {
+          assert.match(err.message, /Ollama API 404/);
+          return true;
+        }
+      );
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
+
+  it('should parse successful response', async () => {
+    const provider = new OllamaProvider({});
+    const mockResponse = {
+      choices: [{ message: { content: 'Hello from Ollama' } }],
+      usage: { prompt_tokens: 12, completion_tokens: 8 },
+      model: 'llama3.1:8b',
+    };
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn(() =>
+      Promise.resolve({ ok: true, json: () => Promise.resolve(mockResponse) })
+    );
+    try {
+      const result = await provider.complete('You are helpful.', 'Say hello');
+      assert.equal(result.text, 'Hello from Ollama');
+      assert.equal(result.usage.inputTokens, 12);
+      assert.equal(result.usage.outputTokens, 8);
+      assert.equal(result.model, 'llama3.1:8b');
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
+
+  it('should send correct request format', async () => {
+    const provider = new OllamaProvider({ model: 'qwen2.5:14b', baseUrl: 'http://myhost:11434' });
+    let capturedUrl, capturedOpts;
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn((url, opts) => {
+      capturedUrl = url;
+      capturedOpts = opts;
+      return Promise.resolve({
+        ok: true,
+        json: () => Promise.resolve({
+          choices: [{ message: { content: 'ok' } }],
+          usage: { prompt_tokens: 1, completion_tokens: 1 },
+          model: 'qwen2.5:14b',
+        }),
+      });
+    });
+    try {
+      await provider.complete('system prompt', 'user message', { maxTokens: 2048 });
+      assert.equal(capturedUrl, 'http://myhost:11434/v1/chat/completions');
+      assert.equal(capturedOpts.method, 'POST');
+      const headers = capturedOpts.headers;
+      assert.equal(headers['Content-Type'], 'application/json');
+      assert.equal(headers['Authorization'], undefined);
+      const body = JSON.parse(capturedOpts.body);
+      assert.equal(body.model, 'qwen2.5:14b');
+      assert.equal(body.max_tokens, 2048);
+      assert.equal(body.messages[0].role, 'system');
+      assert.equal(body.messages[0].content, 'system prompt');
+      assert.equal(body.messages[1].role, 'user');
+      assert.equal(body.messages[1].content, 'user message');
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
+
+  it('should handle empty response gracefully', async () => {
+    const provider = new OllamaProvider({});
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn(() =>
+      Promise.resolve({
+        ok: true,
+        json: () => Promise.resolve({ choices: [], usage: {} }),
+      })
+    );
+    try {
+      const result = await provider.complete('sys', 'user');
+      assert.equal(result.text, '');
+      assert.equal(result.usage.inputTokens, 0);
+      assert.equal(result.usage.outputTokens, 0);
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
+
+  it('should use longer default timeout than cloud providers', async () => {
+    const provider = new OllamaProvider({});
+    let capturedOpts;
+    const originalFetch = globalThis.fetch;
+    globalThis.fetch = mock.fn((url, opts) => {
+      capturedOpts = opts;
+      return Promise.resolve({
+        ok: true,
+        json: () => Promise.resolve({
+          choices: [{ message: { content: 'ok' } }],
+          usage: { prompt_tokens: 1, completion_tokens: 1 },
+        }),
+      });
+    });
+    try {
+      await provider.complete('sys', 'user');
+      assert.ok(capturedOpts.signal, 'Should have an abort signal');
+    } finally {
+      globalThis.fetch = originalFetch;
+    }
+  });
+});
+
+// ─── Factory Tests ───
+
+describe('createLLMProvider — ollama', () => {
+  it('should create OllamaProvider for provider=ollama', () => {
+    const provider = createLLMProvider({ provider: 'ollama', apiKey: null, model: null });
+    assert.ok(provider instanceof OllamaProvider);
+    assert.equal(provider.name, 'ollama');
+    assert.equal(provider.isConfigured, true);
+  });
+
+  it('should be case-insensitive', () => {
+    const provider = createLLMProvider({ provider: 'Ollama', apiKey: null, model: null });
+    assert.ok(provider instanceof OllamaProvider);
+  });
+
+  it('should pass baseUrl from config', () => {
+    const provider = createLLMProvider({ provider: 'ollama', apiKey: null, model: 'mistral:7b', baseUrl: 'http://gpu-box:11434' });
+    assert.ok(provider instanceof OllamaProvider);
+    assert.equal(provider.baseUrl, 'http://gpu-box:11434');
+    assert.equal(provider.model, 'mistral:7b');
+  });
+});