The ollama 0.14-rc2 release introduces experimental functionality allowing LLMs to use tools like bash and web searching on your system, with safeguards like interactive approval and command allow/denylists.
This article details how to build powerful, local AI automations using n8n, the Model Context Protocol (MCP), and Ollama, aiming to replace fragile scripts and expensive cloud-based APIs. These tools work together to automate tasks like log triage, data quality monitoring, dataset labeling, research brief updates, incident postmortems, contract review, and code review – all while keeping data and processing local for enhanced control and efficiency.
**Key Points:**
* **Local Focus:** The system prioritizes running LLMs locally for speed, cost-effectiveness, and data privacy.
* **Component Roles:** n8n orchestrates workflows, MCP constrains tool usage, and Ollama provides reasoning capabilities.
* **Automation Examples:** The article showcases several practical automation examples across various domains, from DevOps to legal compliance.
* **Controlled Access:** MCP limits the model's access to only necessary tools and data, enhancing security and reliability.
* **Closed-Loop Systems:** Many automations incorporate feedback loops for continuous improvement and reduced human intervention.
**Key Points:**
* **Local Focus:** The system prioritizes running LLMs locally for speed, cost-effectiveness, and data privacy.
* **Component Roles:** n8n orchestrates workflows, MCP constrains tool usage, and Ollama provides reasoning capabilities.
* **Automation Examples:** The article showcases several practical automation examples across various domains, from DevOps to legal compliance.
* **Controlled Access:** MCP limits the model's access to only necessary tools and data, enhancing security and reliability.
* **Closed-Loop Systems:** Many automations incorporate feedback loops for continuous improvement and reduced human intervention.
This article details how to build a 100% local MCP (Model Context Protocol) client using LlamaIndex, Ollama, and LightningAI. It provides a code walkthrough and explanation of the process, including setting up an SQLite MCP server and a locally served LLM.
The series of articles by Adam Conway discusses how the author replaced cloud-based smart assistants like Alexa with a local large language model (LLM) integrated into Home Assistant, enabling more complex and private home automations.
1. **Use a Local LLM**: Set up an LLM (like Qwen) locally using tools such as Ollama and OpenWeb UI.
2. **Integrate with Home Assistant**:
- Enable Ollama integration in Home Assistant.
- Configure the IP and port of the LLM server.
- Select the desired model for use within Home Assistant.
3. **Voice Processing Tools**:
- Use **Whisper** for speech-to-text transcription.
- Use **Piper** for text-to-speech synthesis.
4. **Smart Home Automation**:
- Automate complex tasks like turning off lights and smart plugs with voice commands.
- Use data from IP cameras (via Frigate) to control external lighting based on presence.
5. **Hardware Recommendations**:
- Use Home Assistant Voice Preview speaker or DIY alternatives using ESP32 or repurposed microphones.
1. **Use a Local LLM**: Set up an LLM (like Qwen) locally using tools such as Ollama and OpenWeb UI.
2. **Integrate with Home Assistant**:
- Enable Ollama integration in Home Assistant.
- Configure the IP and port of the LLM server.
- Select the desired model for use within Home Assistant.
3. **Voice Processing Tools**:
- Use **Whisper** for speech-to-text transcription.
- Use **Piper** for text-to-speech synthesis.
4. **Smart Home Automation**:
- Automate complex tasks like turning off lights and smart plugs with voice commands.
- Use data from IP cameras (via Frigate) to control external lighting based on presence.
5. **Hardware Recommendations**:
- Use Home Assistant Voice Preview speaker or DIY alternatives using ESP32 or repurposed microphones.
A tutorial on building a private, offline Retrieval Augmented Generation (RAG) system using Ollama for embeddings and language generation, and FAISS for vector storage, ensuring data privacy and control.
1. **Document Loader:** Extracts text from various file formats (PDF, Markdown, HTML) while preserving metadata like source and page numbers for accurate citations.
2. **Text Chunker:** Splits documents into smaller text segments (chunks) to manage token limits and improve retrieval accuracy. It uses overlapping and sentence boundary detection to maintain context.
3. **Embedder:** Converts text chunks into numerical vectors (embeddings) using the `nomic-embed-text` model via Ollama, which runs locally without internet access.
4. **Vector Database:** Stores the embeddings using FAISS (Facebook AI Similarity Search) for fast similarity search. It uses cosine similarity for accurate retrieval and saves the database to disk for quick loading in future sessions.
5. **Large Language Model (LLM):** Generates answers using the `llama3.2` model via Ollama, also running locally. It takes the retrieved context and the user's question to produce a response with citations.
6. **RAG System Orchestrator:** Coordinates the entire workflow, managing the ingestion of documents (loading, chunking, embedding, storing) and the querying process (retrieving relevant chunks, generating answers).
1. **Document Loader:** Extracts text from various file formats (PDF, Markdown, HTML) while preserving metadata like source and page numbers for accurate citations.
2. **Text Chunker:** Splits documents into smaller text segments (chunks) to manage token limits and improve retrieval accuracy. It uses overlapping and sentence boundary detection to maintain context.
3. **Embedder:** Converts text chunks into numerical vectors (embeddings) using the `nomic-embed-text` model via Ollama, which runs locally without internet access.
4. **Vector Database:** Stores the embeddings using FAISS (Facebook AI Similarity Search) for fast similarity search. It uses cosine similarity for accurate retrieval and saves the database to disk for quick loading in future sessions.
5. **Large Language Model (LLM):** Generates answers using the `llama3.2` model via Ollama, also running locally. It takes the retrieved context and the user's question to produce a response with citations.
6. **RAG System Orchestrator:** Coordinates the entire workflow, managing the ingestion of documents (loading, chunking, embedding, storing) and the querying process (retrieving relevant chunks, generating answers).
This article details how the author successfully ran OpenAI's Codex CLI against a gpt-oss:120b model hosted on an NVIDIA DGX Spark, accessed through a Tailscale network. It covers the setup of Tailscale, Ollama configuration, and the process of running the Codex CLI with the remote model, including building a Space Invaders game.
Local Micro-Agents That Observe, Log and React. Build powerful micro-agents that observe your digital world, remember what matters, and react intelligently—all while keeping your data 100% private and secure.
Learn to deploy your own local LLM service using Docker containers for maximum security and control, whether you're running on CPU, NVIDIA GPU or AMD GPU.
Ollama has partnered with NVIDIA to optimize performance on the new NVIDIA DGX Spark, powered by the GB10 Grace Blackwell Superchip, enabling fast prototyping and running of local language models.
An encyclopedia where everything can be an article, and every article is generated on the spot. Articles are often full of hallucinations and nonsense, especially with lower parameter models. The project uses Ollama and Go to generate content.
