Ollama Models

You are scraping the Ollama model registry search results page.

Step 1 - Search results page:
The page shows a list of model cards from the Ollama model registry.
Each card displays a model's name, description, capabilities/tags, download count, and when it was last updated.

Extract ALL models visible on the page. Do NOT follow pagination.
Do NOT put any links in next_urls or detail_urls.

For each model, extract the data in this exact JSON format:
{
  "name": "SonofGod/Apex",
  "description": "Adaptive Autonomy & Esoteric Synthesis",
  "tags": ["tools"],
  "pulls": "74",
  "tag_count": 1,
  "updated": "11 months ago"
}

Another example:
{
  "name": "cosmonautsdev/Cosm_v1",
  "description": "Full-stack AI model for rapid app development. Optimized for Python (Django)",
  "tags": [],
  "pulls": "205",
  "tag_count": 1,
  "updated": "7 months ago"
}

Rules:
- Extract the model name exactly as shown (e.g. "username/modelname")
- Include the full description text
- Tags are capability labels like "tools", "vision", "embedding" etc. Put them in a list.
- pulls is the download/pull count as a string
- tag_count is the number of available tags/versions
- updated is the relative time string (e.g. "2 months ago", "1 year ago")
- If a field is not visible, use an empty string "" or empty list []