We work with search providers and ensure that we have zero data retention policies in place.

The search results are yours to own and use. You are free to do what you want with it. Of course you are bound by local laws of the legal jurisdiction you are in.

You can say you're training an AI model and do whatever you want with it.

You should ask if search results are even copyrightable, if they are just a list of links.

It is strange to launch this type of functionality with not even a privacy policy in place.

It makes me wonder if they’ve partnered with another of their VC’s peers who’s recently had a cash injection, and they’re being used as a design partner/customer story.

Exa would be my bet. YC backed them early, and they’ve also just closed a $85M Series B. Bing would be too expensive to run freely without Microsoft partnership.

Get on that privacy notice soon, Ollama. You’re HQ’d in CA, you’re definitely subject to CCPA. (You don’t need revenue to be subject to this, just being a data controller for 50,000 Californian residents is enough.)

https://oag.ca.gov/privacy/ccpa

I can imagine the reaction if it turns out the zero-retention provider backing them ended up being Alibaba.

There are very few recently launched pure open source projects these days (most are at least running donation-ware models or funded by corporate backers), none in the AI space that I'm aware of.

They are former Docker employees running Docker playbook.

Perhaps not quite solving your problem, but I have a handful of domain-specific Google CSE (Custom Search Engine) that limit the results to predefined websites. I summon them from Alfred with short keywords when I'm doing interest-specific searches. https://blog.gingerbeardman.com/2021/04/20/interest-specific...

YaCy (https://yacy.net) can do all this I think. Cloudflare might block you IP pretty soon though if you try to crawl.

Have you ever looked at Common Crawl dumps? I did a bit of data mining and holy cow is 99.99% of the web crap. Spam, porn, ads, flame wars, random blogs by angsty teens... I understand it has historical and cultural value — and maybe literary value, in a Douglas Coupland kind of way — but for my purposes, there was very little here that I considered of interest.

Which was very encouraging to me, because it implies that indexing the Actually Important Web Pages might even be possible for a single person on their laptop.

Wikipedia, for comparison, is only ~20GB compressed. (And even most of that is not relevant to my interests, e.g. the Wikipedia articles related to stuff I'd ever ask about are probably ~200MB tops.)

Have you ever tried https://marginalia-search.com ? I love it.

Kind of. I made ainews247.org that crawls certain sites and filters content so it's AI specific and valuable. I think it's a really good idea.

You could take a look at the leaked Yandex source code from a few years ago. I'd believe their architecture should be decent enough.

With LLMs why do you even need a mini-Google?

Fair question. Some of the supported models are large and wouldn't fit on most local devices. This is just the beginning, and Ollama does not need to exclude cloud hosted frontier models either with the relationship we've built with the model providers. We just have to be mindful and understand that Ollama stands with developers, and solve the needs.

https://ollama.com/cloud

You make an account to use their hosted models AND to have them available via the Ollama API LOCALLY. I'm spending $100 on Claude and $200 on GPT5, so $20 bucks is NOTHING and totally worth having access to:

Qwen3 235b

Deepseek 3.1 671b (thinking and non thinking)

Llama 3.1 405b

GPT OSS 120b

Those are hardly "small inferior models".

What is really cool is that you can set Codex up to use Ollama's API and then have it run tools on different models.

For models you can't run locally like gpt-oss-120b, deepseek or qwen3-coder 480b. And a way for them to monetize the success of Ollama.

Yeah it's been a steady pivot to profitable features. Wonderful to see them build a reputation through FOSS and codebase from free labor to then cash in.

a lot of "local" models are still very large to download and slow to run on regular hardware. I think it's great to have a way to evaluate them cheaply in the cloud before deciding to pull down the model to run locally.

At some level it's also more of a principle that I could run something locally that matters rather than actually doing it. I don't want to become dependent on technology that someone could take away from me.

How do you meaningfully use it without using scraping APIs? Aren't the official apis severely limited?

We did consider building functionality into Ollama that would go fetch search results and website contents using a headless browser or similar. However we had a lot of worries about result quality and also IP blocking from Ollama creating crawler-like behavior. Having a hosted API felt like a fast path to get results into users' context window, but we are still exploring the local option. Ideally you'd be able to stay fully local if you want to (even when using capabilities like search)

Their GUI is closed-source. If someone wants an easy to use & easy to setup app, may as well use LMStudio, which doesn't try to pretend to be OSS. Or use ramalama which is basically just containerizing LLMs and the relevant bits, pretty damn similar to ollama. Or just go back to "basics" and use llama.cpp or vllm.

Their posture has continually been getting worse and worse. It's deceptive and I've expunged it from all my systems

To provide additional features or using Ollama's cloud hosted models, you can signup for an Ollama account.

For starter, this is completely optional. It can be completely local too for you to publish your own models to ollama.com that you can share with others.

Sorry about this. We are working really hard on providing a usage based pricing.

During the preview period we want to start offering a $20 / month plan tailored for individuals - and we are monitoring the usage and making changes as people hit rate limits so we can satisfy most use cases, and be generous.

That's the essence of these services, they never explicitly mention the quota, or secretly lowers it at some point.

Hey! Author of the blogpost and I also work on Ollama's tool calling. There has been a big push on tool calling over the last year to improve the parsing. What's the issues you're running into with local tool use? What models are you using?

You can use solr, very good full text search and it has an mcp integration. That’s sufficient on its own and straightforward to setup:

https://github.com/mjochum64/mcp-solr-search

A slightly heavier lift, but only slightly, would be to also use solr to also store a vectorized version of your docs and simultaneously do vector similarity search, solr has built in knn support fort it. Pretty good combo to get good quality with both semantic and full-text search.

Though I’m not sure if it would be relatively similar work to do solr w/ chromadb, for the vector portion, and marry the result stewards via llm pixie dust (“you are the helpful officiator of a semantic full-text matrimonial ceremony” etc). Also not sure the relative strengths of chromadb vs solr on that- maybe scales better for larger vector stores?

docling might be a good way to go here. Or consider one of the existing full text search engines like Typesense.

Based on the fact that there are very few up-to-date English-language search indexes (Google, Bing, and Brave if you count it), it must be incredibly costly. I doubt they are maintaining their own.

AgenticSeek, or you can get pretty far with local qwen and Playwright-Stealth or SeleniumBase integrated directly into your Chrome (running with Chrome DevTools Protocol enabled).

sounds like a good way to get your IP flagged by cloudflare

No, Ollama is it's own project and separate. You can check it out via GitHub

https://github.com/ollama/ollama

How is that any different than someone installing an ad blocker in their browser? Arguably ad blocker is much simpler technology than running a local LLM and has been available for years now. And yet Google’s ad revenue seems to have remained unaffected.

They'll have to squeeze it all from Youtube!

They can always pivot to their Search-via-API business :)

It takes lots of servers to build a search engine index, and there’s nothing to indicate that this will change in the near future.

There are millions of websites, and a local LLM cannot scrape all of them to make sense of them. Think about it. OpenAI can do it because they spend millions to train its systems.

Many sites have hidden sitemaps that cannot be found unless submitted to google directly. (Not even listed in robots txt most of the time). There is no way a local LLM can keep up with up to date internet.

google.com/sorry

Hi - author of the post. Yes it does! The "build a search agent" example can be used with a local model. I'd recommend trying qwen3 or gpt-oss

I don't think ollama officially supports any proper tool use via api.

I was pleasantly surprised on the model improvements when testing this feature.

For smaller models, it can augment it with the latest data by fetching it from the web, solving the problem of smaller models lacking specific knowledge.

For larger models, it can start functioning as deep research.

Taviy gives you 1k free requests a month.

Even with heavy ai usage I'm only at like 400/1000 for the month

Does anyone know if the python and JavaScript examples on the blog work without an Ollama Account?

they had at some point start earning money.

That's what i have together with open webui and gpt-oss-120b. it works reasonably well. But sometimes the searches are slow.

I haven't tried SearXNG personally. How does it compare to Ollama's web search in terms of the search content returned?