docs: update provider docs and add changeset for Ollama integration

Author: Dhruv1969Karnwal

.changeset/ollama-integration.md

@@ -0,0 +1,20 @@
+# .changeset/ollama-integration.md
+
+---
+
+"@voltagent/ollama": minor
+"@voltagent/core": patch
+
+---
+
+feat(ollama): Implement Ollama provider integration
+
+This feature allows users to use local Ollama instance as an LLM provider within VoltAgent applications.
+
+- Adds OllamaProvider class implementing LLMProvider interface
+- Implements core text generation functionalities
+- Adds configuration for Ollama endpoint URL
+- Includes example implementation
+- Updates provider documentation
+
+Fixes #15

website/docs/agents/providers.md

@@ -57,3 +57,90 @@ Always refer to the documentation for the specific provider you are using to kno
 VoltAgent offers built-in providers for various popular services and SDKs, including Vercel AI, Google AI (Gemini/Vertex), Groq, and any OpenAI-compatible API endpoint (via xsAI).
 
 **For detailed information on each available provider, including specific installation, configuration, and usage instructions, please see the [Providers Overview](../providers/overview.md).**
+
+### `@voltagent/ollama` (Ollama AI Provider)
+
+Connects to Ollama and allowing you to use your locally hosted Ollama models with Voltagent.
+
+## Prerequisites
+
+1. [Install Ollama](https://ollama.ai/download) on your machine
+2. Pull the models you want to use:
+   ```bash
+   ollama pull llama3
+   # or any other model you want to use
+   ```
+3. Ensure Ollama is running:
+   ```bash
+   ollama serve
+   ```
+
+## Usage
+
+```typescript
+import { Agent } from "@voltagent/core";
+import { OllamaProvider } from "@voltagent/ollama";
+
+// Initialize the Ollama provider
+const ollama = new OllamaProvider({
+  // Optional: provide a custom URL if Ollama is not running on the default localhost:11434
+  baseUrl: "http://localhost:11434",
+});
+
+// Create an agent using the Ollama provider
+const agent = new Agent({
+  name: "LocalLLMAgent",
+  description: "An agent that uses a local Ollama model",
+  llm: ollama,
+  // Specify the model name - this should match a model you've pulled in Ollama
+  model: "llama3",
+});
+
+// Use the agent
+const response = await agent.generateText("Hello, what can you do for me?");
+console.log(response.text);
+```
+
+## Provider-Specific Options
+
+When calling agent methods, you can pass provider-specific options:
+
+```typescript
+const response = await agent.generateText("What is the capital of France?", {
+  provider: {
+    temperature: 0.7,
+    maxTokens: 500,
+    topP: 0.9,
+    seed: 123,
+  },
+});
+```
+
+## Available Models
+
+The available models depend on what you've pulled to your Ollama instance. Common models include:
+
+- `llama3`
+- `mistral`
+- `gemma`
+- `phi`
+- `llama2`
+- `llama2-uncensored`
+- `orca-mini`
+- `codellama`
+- `stable-code`
+
+Check the [Ollama model library](https://ollama.ai/library) for the full list of available models.
+
+## Example with Streaming
+
+```typescript
+const stream = await agent.streamText("Tell me a story about a robot");
+
+// Process the stream
+for await (const chunk of stream.textStream) {
+  process.stdout.write(chunk);
+}
+```
+
+**Choose @voltagent/ollama when:** You want to use models available through Ollama, which provides a convenient local-first framework for running open-source language models. Ollama is well-suited for tasks that benefit from customizable, offline-friendly models, including text generation, lightweight coding tasks, and privacy-sensitive applications.