Onyx
onyx.app
- Free plan
- Open source
- Windows
- macOS
- Linux
- Web
- Browser extension
- API
- Self-hosted
What Onyx is
Onyx connects enterprise knowledge to provide context for AI answers to people and agents. It searches indexed sources in parallel and returns relevant documents while enforcing the permissions attached to each source. Its connectors sync data in the background and support hybrid semantic and keyword search; Onyx says it offers 50+ connectors. The Business plan includes chat and search, major LLM access, custom agents, MCP/OpenAPI actions, web search, deep research, a code interpreter and image generation. It costs $20 per user per month with annual billing and includes Community Support. Onyx supports APIs for messages, agents, LLM management, actions and user authentication. It can be deployed in a customer's cloud or on bare metal, including on-premise and air-gapped environments. The site says documents and embeddings can stay within the customer's infrastructure, data is not used for training, and source permissions are checked when queries run. Native Windows, macOS and Linux desktop apps connect to an existing Onyx deployment. A free plan and a 14-day trial are listed.
Who it suits
Onyx describes use across company-wide teams, including engineering, sales and support. It may suit organizations seeking enterprise knowledge search and AI answers with source permissions enforced.
Good to know, and what to watch
- Searches indexed sources in parallel with permissions enforced.
- Offers 50+ connectors that sync data continuously.
- Supports customer-cloud, on-premise and air-gapped deployments.
- Free plan and 14-day trial are listed.
- Business costs $20 per user per month with annual billing.
- Desktop apps require an existing Onyx deployment.
- Dedicated Support and an Enterprise SLA are listed for Enterprise.
Plans
Every plan and priceBusiness
Enterprise
Community Edition
Free plan · 14-day trial
Getting started
- Open https://www.onyx.app/
- Choose a Business plan or discuss Enterprise deployment
- Deploy Onyx in your cloud, on bare metal, or on-premise
- Connect tools through available connectors
- Use the web app, API, or a native Windows, macOS, or Linux desktop app
- Connect Slack integrations for responses and agents
Limits to check
Business costs $20 per user per month with annual billing and includes 40+ app connectors plus Community Support. Enterprise adds on-premise and region-specific deployments, custom integrations, dedicated support, and an Enterprise SLA.
The Club's review
Onyx is an enterprise search and agent platform that turns information in company systems into permission-aware AI answers. It is best suited to organizations with many internal sources, strict data controls, and teams in engineering, sales, support, or company-wide operations. Its strongest advantage is deployment and access-control flexibility; its main compromise is that the principal paid plan requires annual per-user billing.
Overview
Onyx indexes connected services and searches them in parallel, returning relevant documents while checking each source's permissions at query time. More than 50 connectors synchronize tools continuously in the background. Hybrid semantic and keyword search means users can look for an idea or an exact term, which is useful when company knowledge is spread across different systems.
The security and deployment model is unusually broad. Documents and embeddings can remain in the customer's infrastructure, and Onyx says customer data is not used for training. Organizations can deploy in their own cloud or on bare metal, including on-premise and fully air-gapped environments. That is a strong fit for regulated or isolated installations. Teams wanting a provider-managed service with little infrastructure responsibility may find these options more operational than necessary.
Key features
Business brings chat and search together with access to major LLMs, custom agents, MCP/OpenAPI actions, web search, deep research, code interpreter, and image generation. The breadth supports both everyday question answering and more involved research or workflow automation, although organizations should assess which capabilities they actually need before paying for every user.
Onyx provides APIs for messages, agents, LLM management, actions, and user authentication, giving engineering teams ways to embed the service or manage it programmatically. Its integrations place responses and agents inside Slack; Microsoft Teams and Zendesk are described as coming soon, so those channels should not be treated as current deployment targets.
A native desktop app for Windows, macOS, and Linux connects to an existing Onyx deployment. This is useful when staff need a desktop entry point while the organization's data and service remain centrally controlled. Web access and API availability also support browser-based and custom applications.
Pricing
Onyx uses a freemium model. The free plan costs nothing and is suitable for evaluating the search experience or starting with a small group, but the published plan information does not define its seat, usage, or feature caps, so it is not a dependable basis for sizing a rollout. A 14-day trial is available for testing the paid offering.
Business — $20 per user per month, billed annually. It is for teams of any size and includes 40+ app connectors plus Community Support. Chat and search, major LLM access, custom agents, MCP/OpenAPI actions, web search, deep research, code interpreter, and image generation make it the practical choice for organizations that want the full general-purpose feature set without negotiating an enterprise contract.
Enterprise — custom pricing. This plan adds OIDC/SAML SSO, on-premise and region-specific deployments, custom integrations, Dedicated Support, and an Enterprise SLA. It fits organizations whose identity, residency, integration, or service-level requirements exceed Business. Smaller teams that do not need those controls are likely better served by Business or the free plan.
Platforms
Onyx is available on the web, through an API, and on Windows, macOS, and Linux desktop apps. Self-hosted deployment is supported, including customer-cloud, bare-metal, on-premise, and fully air-gapped environments. That range covers browser users, developers, and organizations that cannot place data in a shared hosted environment.
Who it's for
Onyx is a strong match for company-wide knowledge programs and for engineering, sales, and support teams that need answers grounded in internal tools. It is especially relevant where source permissions, data residency, or isolated infrastructure are non-negotiable. It is less compelling for a small group seeking a simple standalone chatbot, or for buyers unwilling to manage deployment choices and an annual per-user commitment.
Pros and cons
- Pros: Query-time source permissions reduce the risk of exposing documents users could not otherwise access.
- Pros: More than 50 continuously synced connectors and hybrid search address fragmented enterprise knowledge.
- Pros: Customer-cloud, on-premise, bare-metal, and air-gapped deployments support demanding security environments.
- Pros: APIs, custom agents, and MCP/OpenAPI actions give developers room to build workflows around the service.
- Cons: Business is $20 per user per month only with annual billing, which raises the commitment for organizations that want a month-to-month pilot.
- Cons: Enterprise controls such as SSO, region-specific deployment, dedicated support, and an SLA require custom pricing.
- Cons: Slack is the current named collaboration integration; Microsoft Teams and Zendesk are not yet presented as available destinations.
Alternatives
For a broader directory of comparable products, see AI Enterprise Search Software and AI chatbots and assistants.
- Elastic Search AI is a better fit when you want the free, self-managed Elastic Stack or a usage-based Serverless option.
- Korra suits buyers wanting mobile and browser coverage with a free tier capped at 100MB.
- Fess is an Apache-licensed, free self-hosted search option.
- Viglet Turing ES is another free Apache 2.0 self-hosted choice.
- Dropbox Dash is for teams preferring a paid hosted product with plans for up to 100 users and beyond.
- OpenSearch is appropriate when open-source deployment with no licensing fees is the priority.
- SWIRL AI Search targets organizations prepared for a paid evaluation pilot before an annual contract.
- GoSearch Analytics is an alternative with a $20-per-user-per-month Pro plan and browser-extension access.
Verdict
Choose Onyx if you need one permission-aware search and agent layer across many company systems, especially with customer-controlled or air-gapped deployment. Look elsewhere if you want a simpler hosted chatbot, open-source software with no licensing cost, or a paid plan without an annual per-user commitment.
The verdict
Onyx brings connected company information into AI search and answers, with permissions enforced at query time. Its deployment options include customer infrastructure, while the Business plan starts at $20 per user per month billed annually.
Questions people ask
Does Onyx offer a free plan?
The listed Business plan costs $20 a month per user with annual billing. An Enterprise plan is listed, but its price is not provided.
Which platforms can run Onyx?
Onyx supports API, web, Windows, macOS, Linux, and self-hosted deployments. Native desktop apps connect to an existing Onyx deployment.
Can Onyx run on-premise or air-gapped?
Yes. It supports customer-cloud and bare-metal deployments, including on-premise and fully air-gapped environments.
What integrations are available?
Onyx says it offers 50+ connectors, while Business lists 40+ app connectors. Responses and agents are available within Slack; Microsoft Teams and Zendesk are described as coming soon.
Does Onyx provide APIs?
Yes. APIs support messages, agents, LLM management, actions, and user authentication.
Is Onyx open source?
The listed facts do not identify Onyx as open source.
What the maker publishes
- Actions and MCP
- Five built-in actions cover internal search, web search, Python code execution, image generation and a beta coding agent; additional actions use OpenAPI or MCP with token or OAuth authentication. source
- API
- Nearly every feature is exposed through a REST API using JSON, with Bearer API keys or personal access tokens and an OpenAPI explorer at /api/docs. source
- API permissions
- API keys act as service accounts whose group memberships control chat, search, agent, connector and administrative access; custom groups are Enterprise-only. source
- Chat
- The chat interface supports file and URL context, model selection, citations, projects, sharing and multiple AI actions. source
- Chrome extension
- The Chrome extension provides new-tab search, a sidebar, selected-text queries and keyboard shortcuts, with full-page questioning listed as future functionality. source
- Company founders
- Crunchbase lists Chris Weaver and Yuhong Sun as founders. source
- Company legal name
- Crunchbase identifies the legal entity as DanswerAI Inc., formerly known as DanswerAI and Danswer. source
- Company location and size
- Crunchbase places the company in San Francisco, California, United States and lists 11–50 employees. source
- Connector behavior
- Connectors sync data in the background and support hybrid semantic and keyword search across connected sources. source
- Connectors
- Onyx says it offers 50+ connectors that sync tools continuously and index them with source permissions. source
- Connectors
- Onyx keeps connected application data synchronized and can enforce source permissions in Enterprise Edition; documents can also be pushed through API or file upload. source
- Custom agents
- Users can define agents with natural-language instructions, scoped connector or uploaded knowledge and selected actions, then share them with users or groups. source
- Deep research
- Deep Research runs repeated reasoning, research and action cycles, can take several minutes and may consume more than ten times normal inference tokens. source
- Deployment
- Onyx supports deployments in a customer's cloud or on bare metal, including on-premise and fully air-gapped deployments. source
- Deployment
- Onyx can run as self-hosted Standard or Lite, in Docker Compose, Kubernetes/Helm or cloud infrastructure, and is also offered as managed Onyx Cloud. source
- Desktop apps
- Official documentation provides downloads for Windows EXE, macOS DMG and Linux DEB, with additional MSI, RPM and tarball formats via GitHub Releases. source
- Desktop platforms
- The native desktop app is available for Windows, macOS, and Linux and connects to an existing Onyx deployment. source
- Developer API
- Onyx supports APIs for messages, agents, LLM management, actions, and user authentication. source
- Documentation depth
- The documentation includes feature guides, API authentication and reference guidance, Docker, Helm, Terraform, cloud deployment, security hardening, authentication and release notes. source
- Features
- The Business plan lists chat and search, major LLM access, custom agents, MCP/OpenAPI actions, web search, deep research, code interpreter, and image generation. source
- File uploads
- Yes source
- Founded
- G2 lists Onyx AI as serving customers since 2023 and gives 2023 as its founding year. source
- Free plan
- Yes source
- Funding
- Crunchbase describes Onyx as a private seed company with two funding rounds and names First Round Capital and Khosla Ventures among investors. source
Checked against the maker's pages on 20 Oct 2026.