Ollama Integration

How TotalApp integrates with Ollama for fully local, private AI processing — and where to configure it.

Ollama Support in TotalApp

TotalApp integrates with Ollama, a local model runner that keeps both text generation and embedding computation entirely on your own machine. Nothing in your knowledge library — documents, notes, or internal material — ever has to leave your computer to be processed. The Node.js server has no network path into your Ollama instance; the browser talks directly to your local 127.0.0.1:11434 endpoint.

Embedding models are a specific category of Ollama model: instead of generating chat responses, they convert text into numeric vectors used to rank and retrieve related content — powering features like Semantic Search and automatic knowledge attachment in Writer Tools.

Where to configure it

Go to Settings → Agentic. From there you can check your Ollama connection status, select an already-installed model, or download a new one from the built-in model list — for either the Writer Engine or the Embedding Engine. See Agentic AI Settings for the full settings reference, Local Language Models for the text-generation model catalog, Embedded Models for the embedding model catalog and download instructions, and Embedding Model for how the selected model is used by Semantic Search.

Text Generation vs. Embeddings

Ollama serves two independent roles inside TotalApp, and they are configured separately:

Text Generation (Writer Engine)

When Ollama is selected as the Writer Engine, chat and content-generation models (e.g. llama3, qwen2.5, deepseek-r1) run locally instead of calling the Anthropic API or CLI.

Embeddings (Embedding Engine)

Independently of the Writer Engine, Ollama is the only supported provider for embedding models — used to rank and retrieve knowledge sources for Semantic Search.

You can mix and match freely: generate content via the Anthropic API while still using a local Ollama model purely for embeddings, or vice versa.

FAQ

Does my data ever leave my computer when using Ollama?

No. Both text generation and embedding requests routed through Ollama are sent directly from your browser to your local 127.0.0.1:11434 endpoint — the TotalApp server never sees this traffic.

Do I need Ollama installed to use TotalApp?

No. Ollama is entirely optional. Without it, text generation falls back to your configured Writer Engine (Hosted API or Local CLI), and Semantic Search ranking is simply unavailable until an embedding model is configured.

Where can I find the list of downloadable text-generation models?

See Local Language Models for the full catalog, grouped by what each model is best at.

Where can I find the list of downloadable embedding models?

See Embedded Models for the full catalog, recommended starting points, and download instructions.