Zenolve Feeds
Open-source AI data

Ollama Model Library API

The full Ollama model library as structured JSON - model name, pull count, sizes, capabilities, and last update.

Source: Ollama$0.005 / recordJSON · CSV · ExcelMCP tool for AI agents

What this feed does

Ollama's library is the de-facto leaderboard of local LLM adoption - but it is a web page, not an API. This feed scrapes the full library on demand and returns every model as structured JSON, including its pull count (the demand signal).

Structured, not scraped-by-you

Clean JSON with stable field names - no HTML parsing, no selectors to maintain when the site changes.

Pay only for what you pull

No subscription. $0.005 per record via Apify's metered billing; a typical pull is cents.

Human + agent ready

Call it from cURL, Python, or JS - or hand it to an AI agent as an MCP tool.

Always the latest

Scraped from the official source on every run, so you never ship stale data.

What each record contains

Every result is a JSON object with these fields.

FieldTypeDescription
namestringModel name (e.g. llama3.3, qwen2.5).
pullsstringTotal pull count - the adoption signal.
sizesstringAvailable parameter sizes (first listed).
capabilitiesstringCapability tag (tools, vision, embedding, ...).
tag_countstringNumber of published tags/variants.
updatedstringRelative last-update time.
descriptionstringModel description.
urlstring, URLLink to the model page on ollama.com.

Live sample

The most recent records, straight from the feed. Live sample - fetched at build time

namepullssizesupdated
llama3.1117.2M8b1 year ago
deepseek-r189.8M1.5b1 year ago
nomic-embed-text78.5M2 years ago
llama3.276.6M1b1 year ago
gemma338.6M270m11 months ago
qwen2.534.7M0.5b1 year ago
qwen332.3M0.6b9 months ago
mistral31.1M7b1 year ago

Call it in one line

Runs the actor and returns the dataset. Swap in your Apify token.

cURL - one call, JSON back
curl "https://api.apify.com/v2/acts/zenolvepro~ollama-library/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN"
Python
from apify_client import ApifyClient
client = ApifyClient("YOUR_APIFY_TOKEN")
run = client.actor("zenolvepro/ollama-library").call()
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row)
JavaScript / Node
import { ApifyClient } from 'apify-client';
const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('zenolvepro/ollama-library').call();
const { items } = await client.dataset(run.defaultDatasetId).listItems();
console.log(items);
AI agents - Model Context Protocol (MCP)
https://mcp.apify.com?tools=zenolvepro/ollama-library

Point any MCP-capable agent at this endpoint and the feed becomes a callable tool.

Who uses it

AI VCs, devtool marketers, and ML-infra vendors sizing open-model demand use it for monitoring, benchmarking, alerting, and enrichment.

Pricing

$0.005per record returned · pay-per-result, no subscription

Billed through Apify. A tiny per-run start fee ($0.00005) applies. Only pay for records you actually pull.

FAQ

Is there an official Ollama library API?
No - the library is a web page. This feed returns it as structured JSON at $0.005 per record.
What makes pull counts useful?
They are the closest public proxy for local-LLM adoption - which models developers actually download and run.
How current is it?
Live - each call scrapes ollama.com/library at request time.
What does each record include?
name, pulls, sizes, capabilities, tag_count, updated, description, and url.
Can an AI agent call this directly?
Yes - exposed as an MCP tool for native agent calls.