Ollama

Ollama

Runs open models on your own hardware from a terminal, free and unlimited, fully offline.

Last verified Uncensored AI

What is Ollama?

Ollama is an open-source tool that pulls open-weight models and runs them on your own hardware, driven from a terminal or a small desktop app. It exposes a local HTTP API that most model tooling already understands, which is why a great deal of other software can be pointed at it. Once a model is pulled it runs with no internet connection at all.

At a glance

Key facts about Ollama, last verified 8 Aug 2026
CategoryUncensored AI
Pricing modelFreemium
Starting price$20 per month
Free tierYesRunning models on your own hardware is unlimited and always free. The free tier also includes light cloud usage.
Free trialNo
Card required at signupNo
Account requiredNo
Payment methodsCard
PlatformsWindows, macOS, Linux, API
Generation modesText chat, Text generation
Content filteringNo documented output filter
Age policyNone stated
Position on minorsNot applicableRuns on your hardware and receives no user content.
Data retentionLocal use retains nothing off the machine: models run on your hardware and can run entirely offline. Ollama states that your data is never trained on, and describes zero data retention and no logging on the Team tier. Cloud models are hosted in the United States, Europe and Singapore.source
Trains on your contentNo
Deletion on requestSelf-serve in account settings
Anonymous useSupported
CompanyOllama Inc., United States
Last verifiedby ClickTugger

Product identity

Identity and operator of Ollama
Official nameOllamaVerified
OperatorOllama Inc.Documented
JurisdictionUnited StatesDocumented
Websiteollama.comVerified
Also known asollama.comVerified

Who it is for

People comfortable with a command line who want models running on their own machine, and developers who want a local endpoint to build against. It is the default choice for anyone who wants to script model use rather than click through an interface.

Who it is not for

Anyone who wants a polished chat interface out of the box. Ollama is deliberately plain, and while desktop front-ends exist, the core experience is a terminal. It also produces text only, with no image or video generation.

Pricing

Running models on your own hardware is free and unlimited, permanently, and that is the part most readers of this directory want. The paid tiers buy hosted cloud inference rather than removing a local restriction: Pro is 20 US dollars a month or 200 a year, Max is 100 a month and was paused for new signups when checked, and Team is 25 per seat a month with a five seat minimum.

Yes, and it is unlimited for the thing that matters. Local model use has no cap and needs no account, and the free tier additionally includes a small cloud allowance. Nothing about running a model at home is gated behind payment.

TierPriceIncludes
FreeFree
  • Unlimited local model use on your own hardware
  • Light cloud usage

Local use needs no account. This is the product for most readers.

Pro$20/month
  • Larger cloud models
  • Three cloud models at a time
  • Fifty times the free cloud allowance

Also offered at $200 a year. Cloud only: local use is already unlimited.

Max$100/month
  • Up to ten concurrent cloud models
  • Five times the Pro allowance

Paused for new signups at the time of checking.

Team$25/month
  • Per seat, five seat minimum
  • Shared billing
  • Stated zero data retention and no logging

Refunds: Not applicable to local use, which is free. Cloud subscriptions bill monthly or annually.

Signup friction

No. Local use requires no account at all, so nothing is collected before you run a model. An account and a card are only needed for the optional cloud tiers.

  • No card at signup
  • No account required
  • Anonymous use supported

Platforms

Windows, macOS and Linux, plus a local HTTP API that a large amount of third-party tooling already targets. Cloud models, where used, run in the United States, Europe and Singapore.

  • Windows
  • macOS
  • Linux
  • API

Feature matrix

Absence of evidence is not recorded as absence of a feature. Anything that needs an account to confirm is marked not tested rather than "no".

Features of Ollama
Text chatYesDocumented
Character chatNot offeredDocumented
Image generationNot offeredDocumented
Image to imageNot offeredDocumented
Video generationNot offeredDocumented
VoiceNot offeredDocumented
WebNot offeredDocumented
iOSNot offeredDocumented
AndroidNot offeredDocumented
DesktopYesDocumented
APIYesDocumented
Runs locallyYesDocumented
Long-term memoryNeeds an accountNot tested
Group chatNeeds an accountNot tested
Custom system promptNeeds an accountNot tested
Model selectionNeeds an accountNot tested

Privacy and data

For local use there is nothing to hand anywhere: the model runs on your machine and works offline once pulled, so no prompt reaches a server. Ollama states plainly that your data is never trained on, and describes zero data retention and no logging on the Team tier. Cloud usage is a separate hosted service in named regions, and choosing it means prompts leaving your machine under different terms from the local mode.

RetentionLocal use retains nothing off the machine: models run on your hardware and can run entirely offline. Ollama states that your data is never trained on, and describes zero data retention and no logging on the Team tier. Cloud models are hosted in the United States, Europe and Singapore.source
Trains on your contentNo
DeletionSelf-serve in account settings
Processing locationYour own device for local use
Privacy policyRead the vendor policy

Privacy disclosure

Privacy disclosure by Ollama
Trains on your contentStated: nosourceDocumented
Account deletionA control in the accountDocumented
Deletion control seen workingNeeds an accountNot tested
Chat exportNeeds an accountNot tested
Usable without an accountYesDocumented
Payment methods disclosedYesDocumented

What the vendor lets you know before paying

This measures disclosure, not quality. A good product can document nothing, and that is worth knowing separately from whether it is any good.

Documentation published by Ollama
Price publishedYes, from $20sourceVerified
Operating company namedOllama Inc.Documented
Jurisdiction statedUnited StatesDocumented
Retention period statedNo period givensourceNot documented
Training policy statedYessourceDocumented
Refund terms publishedYesDocumented
Content policy publishedSelf-hosted, not applicableNot documented
Position on minors publishedNot applicableNot documented

Content policy and safety

Ollama documents its output filtering asno documented output filter, and its age policy as none stated.

Ollama does not offer clothing removal from photographs of real people. ClickTug does not list tools that do. Its stated position on the likeness of real people is that it is blocked outright.

What the vendor prohibits

  • Nothing is enforced by the local client, which receives no content

The client applies no filter. What a model refuses is a property of the model you pulled, and the catalogue includes both heavily aligned and permissive ones. The optional cloud tier is a hosted service under separate terms.

Ollama runs on your own hardware, so it receives none of what you generate. ClickTug does not require a minor safety policy or an acceptable use policy from a tool that never receives user content, because there would be nothing for such a policy to govern. Any optional hosted tier is a separate service under its own terms.

ClickTug verdict

The standard way to run an open-weight model on your own hardware, and the right recommendation for anyone who reaches this category through a command line. Local use is free, unlimited and offline, the client is open source, and the local API makes it the piece most other local tooling assumes. The cloud tiers are a genuine hosted service rather than a paywall around the local product, which is an unusually honest structure. The cost is that Ollama is deliberately plain, and choosing a model is entirely your problem.

Strengths

  • Local use is free, unlimited and needs no account
  • Open source, so the client can be audited or built from source
  • Local HTTP API that a large amount of other tooling already targets
  • Runs fully offline once a model is pulled
  • Paid tiers buy hosted capacity rather than removing a limit on local use

It is the least ceremonious way to get a model running locally, it is open source so the client can be inspected or built from source, and the local API means a lot of other software works with it immediately. For anyone scripting rather than chatting it is the obvious pick.

Limitations

  • Terminal first, with no polished chat interface in the core product
  • Text only. No image or video generation
  • Model choice and quality are entirely the user's problem
  • Same hardware constraints as any local runtime
  • The cloud tiers are a hosted service under separate terms

The plainness is real. There is no polished interface, model selection is on you, and the same hardware constraints apply as to any local runtime. If you want something with a chat window and a model picker rather than a terminal, a desktop-first tool will suit you better.

Hands on

Ollama is the piece most other local model tooling assumes is there. Pull a model with one command, and it runs on your machine behind a local HTTP API that a large amount of third-party software already speaks. That combination is why it turns up as a dependency of things that never mention it on their own homepage.

For this directory the relevant property is the same one that makes LM Studio worth listing: there is no service in the middle. No account is required to run a model locally, nothing is sent anywhere, and after the initial download the whole thing works with the network off. Ollama states that your data is never trained on, and the Team tier goes further with zero data retention and no logging, but for local use those commitments matter less than the architecture does. A promise not to retain your prompts is worth something. Never receiving them is worth more.

The pricing structure deserves a note because it is unusually honest for this market. Ollama sells cloud inference at 20 dollars a month for Pro, 100 for Max, and 25 per seat for Team with a five seat minimum. None of it changes anything about local use. Running models on your own hardware stays unlimited and free whether you subscribe or not, which is the opposite of the common pattern where a free tier exists mainly to demonstrate the limits it imposes.

What you are choosing against is convenience. Ollama is plain by design. There is no model recommendation engine, no interface to speak of in the core product, and no help when a model turns out to be a poor fit for what you wanted. Model selection is genuinely your problem, and the difference between a good and a bad choice is larger than most people expect before they try it. If that sounds tiring rather than appealing, LM Studio does the same job with a window and a browser attached.

As with LM Studio, this listing carries ClickTug’s self-hosted exemption: no minor safety policy is required from a client that receives no user content, because there would be nothing for one to govern. The exemption is recorded against the listing rather than applied quietly, and it does not extend to Ollama’s own cloud tiers, which are a hosted service like any other.

The Ollama home page: a line drawing of a llama above the heading Build with open models, on your computer and in the cloud, with a download button and a terminal panel showing the ollama command
Ollama leads with the terminal, which is an accurate promise. The panel on the right is the whole interface for most users.

How Ollama compares

Every figure on the right is computed across the 37 tools ClickTug currently tracks, at the moment this page was built. They are facts about this catalogue rather than about the market, and they move as it grows.

Ollama compared with the ClickTug catalogue
MeasureOllamaAcross ClickTug
Starting price$20$11.99 median of 17 that publish one
Publishes a priceYes26 of 37 tracked
Free tierYes30 of 37 tracked
Names its operating companyYes33 of 37 tracked
States a training positionYes18 of 37 tracked

At $20 a month, Ollama is cheaper than 1 of the 17 tools here that publish a starting price.

Ollama compared with the alternatives

Every cell is read from the same records as each tool's own listing, so nothing here can claim something its page denies. Rows needing an account to confirm are left out entirely rather than guessed at, and live in each listing's feature matrix instead.

Ollama compared with its alternatives on price, features and disclosure
 OllamaLM StudioVenice AI
Starting price$20/moFree$18/mo
Free tierYesYesYes
Card to startNoNoNo
Account requiredNoNoYes
Runs on your hardwareYesYesNo
Image generationNoNoYes
Video generationNoNoYes
VoiceNoYesYes
APIYesYesYes
Mobile appNoNoNo
Trains on your contentStates noStates noNot stated
Account deletionIn accountIn accountBy request
Company namedYesYesYes
Position on minorsN/AN/APublished
  • LM Studio

    Ollama vs LM StudioFreemium

    Choose LM Studio instead if you want a desktop application with a model browser and a chat window rather than a command line.

  • Venice AI

    Ollama vs Venice AIFreemium

    Choose Venice instead if you do not have the hardware and would rather pay a hosted service that documents its retention position.

Quick answers about Ollama

Generated from the same records as the tables above, so an answer here cannot contradict a figure there. Questions only appear when the answer says something specific to this product.

Is Ollama free?
There is a free tier. Running models on your own hardware is unlimited and always free. The free tier also includes light cloud usage.
How much does Ollama cost?
Paid plans start at $20 a month. Every tier and what it includes is listed in the pricing section above, taken from the vendor's own page.
Does Ollama require a credit card to start?
No card is required to reach the free layer. One is needed for any paid plan.
Does Ollama offer refunds?
Not applicable to local use, which is free. Cloud subscriptions bill monthly or annually.
Does Ollama have an app?
No. It runs in a web browser only, which is common for adult products because both app stores restrict them.
Does Ollama have an API?
Yes. An API is documented, so the same model can be called from your own code rather than only through the interface.
Can Ollama run locally?
Yes. It runs models on your own hardware, so prompts and output never leave the machine and the retention question does not arise.
Does Ollama train on my conversations?
The vendor states it does not use user content to train models. That is a written commitment, and rarer in this category than it should be.
Can I delete my Ollama account?
Yes, through a control in the account rather than by asking support.
Can I use Ollama without an account?
Yes. The service can be reached without registering, which almost nothing else in this directory permits.
Is Ollama actually uncensored?
ClickTug records no documented output filter, which is a different claim from having no rules: its terms still prohibit a specific list of conduct, set out above.
What company owns Ollama, and where?
Ollama Inc., United States. Publishing an identifiable counterparty is worth noting: several tools in this directory name no company at all.

Ollama FAQ

Is Ollama free?

Running models on your own hardware is free and unlimited, permanently, with no account required. The free tier also includes a small amount of hosted cloud usage. The paid tiers, Pro at 20 US dollars a month, Max at 100 and Team at 25 per seat, all buy cloud capacity rather than removing any restriction on local use.

Does Ollama work offline?

Yes, completely, once a model has been pulled. That is the practical difference between a local runtime and a hosted service: after the download, no prompt leaves your machine and no connection is needed. Ollama markets this explicitly for work that cannot depend on a network.

Is Ollama uncensored?

The client applies no filter of its own. Refusal behaviour belongs entirely to whichever model you pull, and the catalogue includes both heavily aligned models and permissive ones. This is a real difference from a hosted assistant, where the filter belongs to the provider and can be changed without telling you.

Does Ollama train on my data?

Its privacy statement says plainly that your data is never trained on, and for local use the question is largely moot because prompts never leave your hardware. The Team tier additionally states zero data retention and no logging. Cloud usage on any tier is a hosted service and should be read as such.

What is the difference between Ollama and LM Studio?

They solve the same problem with different manners. Ollama is command-line first, open source, and built to be driven by other software through its local API. LM Studio is a desktop application with a model browser and a chat window. Both run open-weight models on your own hardware and both are free for that, so the choice is mostly about whether you would rather type a command or click.

How we checked Ollama

Checked
8 August 2026
Documents read
3
Claims checked
0
Documentation gaps
3

What we did

  • Read every policy document linked from this page, in full
  • Loaded the pricing page, and re-loaded it in a rendering browser where it was empty
  • Checked the homepage against the pricing and policy pages for disagreement
  • Ran the vendor's own homepage through ClickTug's excluded-terms gate
  • Searched for reporting, incident-database entries and independent testing
  • Captured the interface and reviewed the capture by eye

What we have not done

Everything below needs an account, and ClickTug does not hold one here. These are recorded as not tested rather than guessed at, which is why they are absent from the tables above rather than marked "no".

  • Signup completed
  • Email verification required
  • Card required at signup
  • First response generated
  • Long-term memory
  • Voice feature
  • Character creation
  • Delete-account control
  • Chat export
  • Cancellation flow

Verified ClickTug observed this directly on the vendor's own pages.Documented The vendor states this in a document we have linked.Not documented We read the document that would carry this and it is absent.Not tested Checking this needs an account. We have not held one, so we have not checked.

Sources for this listing

ClickTug last checked these facts on 8 Aug 2026 against 3 vendor sources. Pricing and data policies in this category change within weeks, so treat the date as part of the fact.

Something wrong or out of date? Report this listing. If this listing concerns you personally, use our removal request form.

More uncensored ai