Ollama
Free, open-source tool to run open-weight LLMs locally via CLI, desktop, or API — with an optional paid cloud for larger models
FREEMIUM Verified JUN 23, 2026
Open WebUI
Self-hosted web interface for local and API-based LLMs, free for unlimited users as long as the branding stays intact
Open WebUI is a self-hosted chat interface that connects to Ollama, OpenAI, Anthropic, or any OpenAI-compatible endpoint, with roughly 145,000 GitHub stars and no vendor-managed cloud. Self-hosting is free for an unlimited number of users, but since v0.6.6 (April 2025) the licence forbids removing Open WebUI branding in deployments above 50 end users in a rolling 30-day window unless you buy an enterprise licence, whose price is not published. Best for teams that can run and maintain their own infrastructure.
Verified JUL 18, 2026 OPEN-SOURCE Live
What is Open WebUI?
Open WebUI is a chat interface you run on your own hardware. It talks to local models through Ollama and to hosted ones through OpenAI, Anthropic, or any OpenAI-compatible endpoint, so a single deployment can front both a laptop-sized local model and a frontier API without the user noticing the difference. Installation is one pip command or a Docker container, with no account required. Around the chat sit the features teams usually assemble from three or four separate products: retrieval over uploaded documents with hybrid keyword-plus-vector search across nine supported vector databases, web search with inline citations, Python code execution, speech-to-text and text-to-speech, and image generation. Administration is unusually complete for a free build — roles, groups, granular permissions, SSO via OIDC, LDAP, and SCIM provisioning all ship in the standard release rather than behind a paywall.
The licence is the detail that catches people out, and it is not plain BSD. Everything up to and including v0.6.5 is BSD-3-Clause. From v0.6.6, released in April 2025, the project ships under the Open WebUI License: the same BSD-3-Clause text plus a fourth clause prohibiting the alteration, removal, or obscuring of Open WebUI branding in any deployment or distribution. Three exemptions apply — deployments with 50 or fewer end users in a rolling 30-day period, substantive contributors with written permission, and holders of an executed enterprise licence. GitHub cannot map this to a standard licence and reports it as NOASSERTION, which means the current release is source-available rather than OSI-approved open source. Practically, use is still free for any number of users; what costs money is white-labelling and enterprise-exclusive capabilities such as Terminals, sold through an enterprise licence with no published price. Anyone who objects to the terms can fork from v0.6.5, which the maintainers explicitly invite.
Who is it for?
Open WebUI suits organisations that have a reason to keep conversations inside their own network and the operations capacity to run a service themselves. There is no managed cloud from the vendor, so hosting, upgrades, backups, and GPU capacity are your problem — in exchange, the software imposes no seat cost, no usage metering, and no data-sharing with a SaaS provider. The reported deployments skew large: a German university is described as serving 30,000-plus students and 5,000-plus employees, and a pharmaceutical company as enabling 3,000-plus employees to share 420-plus custom models.
- Regulated and security-conscious teams in healthcare, finance, defence, or the public sector that need on-premise or air-gapped AI and cannot send prompts to a third-party endpoint.
- Platform and infrastructure engineers who want one governed front door over a mix of local and commercial models, with RBAC and SSO already built in.
- Self-hosting enthusiasts already running local models through Ollama or LM Studio who want a shared, multi-user interface instead of a single-machine app.
- Cost-sensitive organisations with many occasional users, where per-seat hosted chat subscriptions scale badly but a single server does not.
- Teams that must not rebrand unless they buy a licence — worth confirming before design work if your rollout will exceed 50 users.
How much does Open WebUI cost?
Starting price: $0 · Free tier: yes · Model: open-source
Pricing verified JUL 18, 2026
Price history tracked from June 2026
| Plan | Price | Includes |
|---|---|---|
| Self-Hosted (Standard Use) | Free | Free for unlimited end users with original branding intact · Install via one pip command or Docker, no account required · RBAC, groups, SSO/OIDC, LDAP, and SCIM included at no cost · RAG, web search, code execution, voice, and image generation included · You pay only for your own servers and any model APIs you connect |
| Enterprise License | Custom | Required to white-label, rebrand, or alter branding above 50 users · Unlocks enterprise-exclusive capabilities such as Terminals · Adds SLA-backed support and long-term-support releases · No public price — sales contact with seat count required · Sold only to registered entities, not to individual users |
What are Open WebUI's key features?
- Connects to Ollama, OpenAI, Anthropic, and any OpenAI-compatible endpoint from one interface
- Retrieval with hybrid BM25 plus vector search, reranking, and nine supported vector databases
- Eight document extraction engines including Tika, Docling, Mistral OCR, and PaddleOCR
- Web search with inline citations, plus Python code execution from the chat
- Voice input and output with speech-to-text and text-to-speech, and image generation via DALL-E, Gemini, or ComfyUI
- Access control with roles, groups, granular permissions, SSO/OIDC, LDAP, and SCIM provisioning
- API keys using bearer-token auth that reach the same endpoints the web UI uses
- Extensible with community-shared prompts, models, tools, and Python function pipelines
What people use Open WebUI for
- 01 Replacing a hosted chat subscription with an internal deployment where conversations never leave your own network
- 02 Giving a team one interface across local models and several paid provider APIs at once
- 03 Building a retrieval assistant over internal documents using the built-in knowledge bases and vector search
- 04 Running AI in air-gapped or data-residency-constrained environments where SaaS chat tools are not permitted
- 05 Comparing two models side by side on the same prompt before committing to a provider
Pros and cons
| Pros | Cons |
|---|---|
| Free for unlimited users with branding intact — one university deployment is reported to serve 30,000-plus students and 5,000-plus staff | Not OSI-approved open source since v0.6.6 — the added branding clause means GitHub classifies the licence as NOASSERTION rather than BSD-3-Clause |
| Enterprise access controls such as RBAC, SSO, LDAP, and SCIM sit in the free build rather than behind a paywall | Removing or altering Open WebUI branding above 50 end users in any rolling 30-day period is a material licence breach without a paid enterprise licence |
| Installs in one command with no account, and supports fully air-gapped operation | No vendor-managed cloud at all: you own hosting, upgrades, backups, scaling, and GPU capacity yourself |
| Active release cadence, with v0.10.2 shipping on 2026-07-01 and commits landing daily | Enterprise licence pricing is unpublished and gated behind a sales conversation requiring a work email and seat count, and is unavailable to individuals |
| Anyone wanting unrestricted terms can still fork from v0.6.5, which remains BSD-3-Clause | Ships no models of its own, so answer quality and running cost depend entirely on the backends you attach |
What are the best Open WebUI alternatives?
LM Studio
Desktop app to discover, download, and run open-source LLMs locally and privately, with an OpenAI-compatible local server
FREE Verified JUN 24, 2026
Dify
Open-source LLM app platform for building agentic workflows, RAG pipelines, and AI agents with a visual no-code builder
FREEMIUM Verified JUN 24, 2026
ChatGPT
Conversational AI assistant by OpenAI with multimodal reasoning, voice, and agent capabilities
FREEMIUM Verified JUN 11, 2026
Flowise
Open-source low-code platform to build LLM apps and AI agents with a drag-and-drop visual flow builder
OPEN-SOURCE Verified JUN 24, 2026
How people make money with Open WebUI
- Offer Open WebUI deployment and hardening for regulated teams — container or Kubernetes rollout, SSO and RBAC wiring, air-gapped networking — billed as a fixed project fee plus a monthly maintenance retainer
- Build domain-specific knowledge bases, tools, and Python function packs, publish them to the community catalogue for reach, then upsell private configuration work as a recurring engagement
- Run managed hosting for small teams that want the interface without the operations load, charging a per-seat monthly rate while keeping each tenant inside the fifty-user branding threshold
Frequently asked questions
Is Open WebUI free?
Yes. Self-hosting is free for an unlimited number of users, and the official documentation lists standard internal use with original branding intact as free. There is no account requirement and no metered usage. You pay only for your own servers and for whatever model APIs you connect. A paid enterprise licence is needed only for white-labelling or enterprise-exclusive features.
Is Open WebUI open source?
It is source-available rather than OSI-approved open source. Code up to and including v0.6.5 is BSD-3-Clause. From v0.6.6 onward the Open WebUI License adds a clause forbidding removal of the branding, and GitHub classifies the result as NOASSERTION rather than a standard licence. Anyone wanting unrestricted terms can fork from v0.6.5, which the maintainers explicitly permit.
What is the 50-user branding rule?
The licence prohibits altering, removing, or obscuring Open WebUI branding in any deployment, with three exceptions: total end users not exceeding fifty within any rolling thirty-day period, substantive contributors holding prior written permission, or holders of an executed enterprise licence. Below fifty users you may rebrand freely. Above that, removing branding without a licence is a material breach.
How much does an Open WebUI enterprise licence cost?
No price is published. The official enterprise page asks organisations to contact sales from a work email with their expected end-user seat count, and states that licences are available only to registered entities rather than individual users. Any per-seat figure you find on third-party hosting or review sites is not an official Open WebUI price.
Does Open WebUI have an API?
Yes. Open WebUI issues API keys as personal access tokens using standard bearer-token authentication, and the documentation states that scripts can call the same endpoints the web UI uses, including chat completions, model listing, file uploads, and retrieval queries. Each key inherits its creator's role and group permissions, and keys can optionally be restricted to specific routes.
Does Open WebUI offer a hosted cloud version?
No. The vendor ships a self-hosted product only, installed with a single pip command or via Docker, and there is no official managed cloud tier. Several third-party providers sell managed Open WebUI hosting, but those are independent services setting their own prices, and they are not operated by the Open WebUI team.
Which models does Open WebUI work with?
It connects to Ollama for local models and to OpenAI, Anthropic, or any OpenAI-compatible endpoint for hosted ones. You can switch models mid-conversation or run two side by side to compare answers. Open WebUI ships no models itself, so output quality and running costs depend entirely on the backends you attach to it.