Ollama

Give an AI node a model that runs on your own computer through Ollama: free, private, and no API key.

What it does

Ollama is a model you plug into the Model slot of an AI node, such as the Tools Agent or Basic LLM Chain. It does not run on its own and has no output. When the AI node runs, it sends its messages to the Ollama app on your computer, at http://localhost:11434, and gets the reply back. Nothing leaves your machine.

• The Model list shows the models installed in your Ollama, read live from it. Use the refresh button next to the list after you pull a new model.
• Temperature is 0 when you leave it empty. Replies are capped at Max Tokens (4096 by default).
• There is no time limit on a request, because a large model can take a while to load the first time.

When to use it

• Private data that must not go to a cloud AI service.
• Many runs without paying per token.
• Open models such as Llama, Mistral, Qwen or Gemma, at the size your computer can handle.

If you can’t install Ollama, use Web LLM, which runs a model inside the browser.

Inputs and settings

设置项	类型	作用
Model 必填（必填）	Model picker	The model to use.
Options	Group	The options to use with the model.
Options › Temperature	Number	The temperature to use for the model.
Options › Max Retries	Number	The maximum number of retries to use for the model. 默认值：2
Options › Max Tokens	Number	The most tokens the model may generate in one reply. Defaults to 4096, which keeps providers from reserving credit for the whole output window. 默认值：4096

Before the list can fill:

1. Install Ollama from ollama.com and pull a model, for example ollama pull llama3.2.
2. Let AWFlow talk to it. Start Ollama with the OLLAMA_ORIGINS variable set to the extension:
   • Chrome: OLLAMA_ORIGINS=chrome-extension://linlkeaipfpnhddjkpcbmldionajfifa ollama serve
   • Firefox: OLLAMA_ORIGINS=moz-extension://* ollama serve

When you publish or share the workflow, Model is cleared, so whoever installs it picks a model from their own Ollama.

Outputs

This node has no output port. It plugs into a slot on another node instead of passing items along.

• Model slot of Tools Agent, Q&A Agent, RAG Agent, Basic LLM Chain, Indexer, Text Classifier, Sentiment Analysis, Summarization, Information Extractor, Guardrails, Auto-fixing Output Parser

The AI node you connect it to produces the output. Ollama only decides which model answers.

Dependencies and credentials

None. This node needs no connection and has no slots for other nodes.

Example workflow

1. A Context Menu trigger on selected text.
2. Summarization with Text {{ $input.selectionText }} and Max words 60.
3. Ollama in its Model slot, with Model llama3.2:latest picked from the list.
4. Clipboard Write with Text {{ $input.summary }}.

Common issues

• “Could not reach Ollama at localhost:11434. Make sure it is running and allows this extension (OLLAMA_ORIGINS).” Start Ollama, and restart it with OLLAMA_ORIGINS set as shown above.
• The list says “No models available”. Ollama runs but has no models. Pull one (ollama pull llama3.2), then press the refresh button.
• “Model parameter is required and cannot be empty”. Pick a model in Model.
• The first run is very slow. Ollama loads the model into memory on first use. Later runs are faster. A smaller model helps on a laptop.

Related nodes

• Ollama Embeddings: Ollama vectors for a knowledge base.
• Web LLM: a local model inside the browser, nothing to install.
• Tools Agent: an agent that can use tools and memory.