Ollama runs open models such as Llama, Mistral and Qwen on your own machine or server. Connected to Kriti AI, it lets you generate WordPress drafts without an API key and without sending prompts to a third party.
1. Install Ollama and pull a model
Install Ollama from ollama.com, then download a model, for example ollama pull llama3.1. Ollama listens on http://localhost:11434 by default.
2. Connect Kriti AI
In WordPress, open Kriti AI → Settings, enable the Ollama provider, set the endpoint to your Ollama address and choose the model you pulled.
3. Generate a draft
Open Kriti AI → Generate, pick Ollama as the provider, write a prompt and click Generate. The result is saved as a draft post for you to review.
Things to know
- WordPress must be able to reach the Ollama endpoint. If WordPress runs on a remote host, run Ollama on that server or on a machine it can reach.
- Local models are slower than hosted APIs on most laptops. Smaller models (7–8B parameters) are a good place to start.
- Kriti AI’s metrics dashboard still records requests and generation time for local models.