Onyx
Permission-aware AI search for your company’s scattered knowledge and connected tools.
Key Takeaways
- What it is
- Permission-aware AI search for your company’s scattered knowledge and connected tools.
- Pricing
- $20 per user per month
- Best for
- Permission-Aware Enterprise Search
- Indexed Connectors
- Deep Research
- Rating
- 4.4/5
What is Onyx?
Onyx gives teams one place to search company knowledge, chat with connected data, and work with AI agents. It indexes information from tools such as Slack, Google Drive, GitHub, Confluence, Jira, and Salesforce, then returns answers with citations. Teams can run it in managed cloud environments or deploy the open-source edition on their own infrastructure.
Onyx keeps source permissions attached to indexed documents, so users see only information they could already access in the original application. This permission-aware search across continuously synced sources is the clearest reason to consider Onyx for internal AI.
Onyx fits engineering, sales, support, and company-wide knowledge programs that need more control than a standalone chatbot provides. It is less suitable for a person seeking a lightweight personal writing tool, since setup, connectors, permissions, and deployment are built around organizational use. Its strength is turning scattered company data into a governed AI search layer.
Who is Using Onyx?
Software Engineers
Code, pull requests, deployment documents, and team discussions become searchable through one AI interface.
Sales Representatives
Product updates and customer conversations are easier to find when preparing for calls or answering account questions.
Support Specialists
Grounded answers from internal product knowledge help support teams respond with more confidence.
IT and Security Administrators
Deployment controls, identity integration, rate limits, and connector permissions help govern organizational AI use.
AI Platform Developers
REST APIs, an MCP server, and a connector framework support embedded search and custom data sources.
Uncommon Use Cases
Running an internal AI system in an air-gapped environment; building a customer support agent inside another product
Key Features
Rating Breakdown
Pricing Plans
Business
$20 per user per month
- For teams of any size
- Price shown is for annual billing
- Includes chat and search, major LLM access, custom AI agents, MCP and OpenAPI actions, 40+ app connectors, web search, Deep Research, code interpreter, image generation, Slack integration, developer APIs, and usage dashboards
- Includes basic authentication, Google OAuth, role-based access control, permission inheritance, and secret encryption
- Community support
Enterprise
- Everything in Business
- OIDC and SAML SSO
- On-premises and region-specific deployments
- Custom integrations, white-labeling, hook extensions, data exports, volume discounts, invoice billing, dedicated support, and an Enterprise SLA
Which Onyx plan should you pick?
Start with Business for a managed cloud deployment, and choose Enterprise when you need advanced identity, deployment, permission, support, or integration controls.
Business
Good for teams that need managed cloud access to chat, search, agents, connectors, APIs, and usage dashboards.
Enterprise
Good for large or regulated organizations that need SSO, on-premises or region-specific deployment, custom integrations, dedicated support, and an Enterprise SLA.
Onyx Use Cases
Company Knowledge Search
A single search experience brings together answers from workplace tools that normally require separate searches.
Outcome: Cited answers grounded in connected company information.
Engineering Context Retrieval
Engineering teams can connect code repositories, issue trackers, documentation, and team discussions before making technical decisions.
Outcome: Faster access to relevant technical context.
Support Answering
Support teams can find product guidance and internal answers across their approved knowledge sources.
Outcome: More consistent responses based on company knowledge.
Research Across Internal Sources
Deep Research examines information across the knowledge base when a question needs more than one document.
Outcome: Longer, cited analysis from multiple internal sources.
Custom AI Agents
Purpose-built agents combine company data with tools and actions for repeatable team workflows.
Outcome: Agents that can answer questions and take approved actions.
Workflow Ideas with Onyx
Launch Readiness Review
- 1Connect project updates from Slack, documentation, and shared drives.
- 2Ask Onyx for open launch risks and unresolved decisions.
- 3Share a cited briefing with the launch team.
Engineering Incident Review
- 1Search code changes, issue history, incident notes, and team discussions.
- 2Ask for the likely cause and supporting evidence.
- 3Create a follow-up summary through an agent action.
Support Knowledge Workflow
- 1Index product documents and internal help content.
- 2Ask Onyx to draft an answer grounded in approved sources.
- 3Review citations before sending the response.
How Onyx fits into your stack
Company Knowledge Layer
Onyx sits above workplace systems such as Slack, Google Drive, GitHub, Confluence, Jira, and Salesforce. Its connectors continuously index content and related metadata.
AI Chat and Search
The Onyx workspace provides search, chat, citations, web search, Deep Research, and other AI features through one organizational interface.
Agents and Actions
Agents can draw on connected knowledge and call tools through MCP or OpenAPI actions. This gives teams a path from answering questions to completing approved work.
Developer Platform
The REST API, MCP server, and connector framework support embedded experiences, custom sources, and applications built around Onyx.
Infrastructure and Governance
Teams can run Onyx in managed cloud, a VPC, on-premises infrastructure, or an air-gapped environment. Enterprise controls include SSO, role-based access, rate limits, and audit trails.
Pros
- +Permission controls: source-level access rules carry into AI search.
- +Deployment choice: run Onyx in managed cloud, your VPC, on-premises infrastructure, or an air-gapped environment.
- +Open source: the community edition lets teams inspect and customize the application.
- +Connector coverage: common sources include Slack, Google Drive, GitHub, Confluence, Jira, Salesforce, and more.
- +Model flexibility: route work to different providers or use local models.
Cons
- −Setup effort: connecting sources and managing deployment takes more work than using a consumer chatbot.
- −Enterprise dependencies: permission syncing, SSO, and several governance features require Enterprise.
- −Team pricing: the managed Business plan costs $20 per user per month.
Limitations & when to consider alternatives
Known limitations
- Permission Sync Requires Enterprise. The documentation states that syncing document permissions from connected sources requires the Enterprise Edition.
- Enterprise Features Are Tiered. OIDC and SAML SSO, on-premises deployment, custom integrations, data exports, dedicated support, and an Enterprise SLA are listed as Enterprise additions.
- Self-Hosting Requires Infrastructure. The self-hosted option keeps data in your environment, but your team must provide and manage the required infrastructure.
- Connector Coverage Varies. Onyx supports many connectors, but each source has its own setup and synchronization behavior.
Look elsewhere when
- You want a personal AI writing tool rather than company search and governed agents.
- Your team needs a native mobile app.
- Your organization cannot support connector setup or self-hosted infrastructure.
- You need AI model fine-tuning on your own training data.
How to Use Onyx
- 1
Choose a deployment
Start with Onyx Cloud for a managed setup, or follow the self-host guide when your data must remain in your infrastructure.
- 2
Connect your sources
Add supported applications such as Slack, Google Drive, GitHub, Confluence, or Jira so Onyx can index the information your team uses.
- 3
Set access controls
Configure identity, roles, and connector permissions so people receive answers from sources they are authorized to view.
- 4
Ask grounded questions
Search for company information or use chat to receive answers with citations from connected sources.
- 5
Build agents and actions
Create custom agents that combine internal knowledge with MCP or OpenAPI actions for repeatable work.
- 6
Measure adoption
Review query history, usage dashboards, feedback, and audit information to see how teams use AI.
Who is Onyx For?
Engineering Teams
Internal search across code, issues, documentation, and project conversations helps engineers work with fuller context.
Sales Organizations
Connected customer and product information gives sales teams faster access to past conversations and updates.
Customer Support Teams
Support staff can find answers grounded in product knowledge and internal documentation.
Security-Conscious Enterprises
Self-hosting, access controls, audit features, and air-gapped deployment support stricter data requirements.
Open-Source AI Builders
The open-source application and developer APIs provide a base for custom AI search and agent products.
Frequently Asked Questions
Does Onyx have a free option?
Yes. Onyx offers an open-source community edition that can be self-hosted for free. The managed Business plan also includes a free trial.
How much does Onyx cost?
The Business plan costs $20 per user per month. Enterprise uses custom pricing.
Which tools can Onyx connect to?
Onyx lists more than 40 connectors, including Slack, Google Drive, GitHub, Confluence, Jira, Salesforce, SharePoint, Notion, Gmail, HubSpot, and Zendesk.
Can Onyx run inside my network?
Yes. Onyx supports deployment in a VPC, on-premises infrastructure, bare metal, and air-gapped environments.
Does Onyx support SSO?
Enterprise includes OIDC and SAML SSO. Onyx also documents integrations with identity providers such as Okta and Azure AD.
Can developers access Onyx programmatically?
Yes. Onyx provides a REST API, an MCP server, and a connector framework for custom data sources and applications.
Community reviews
Be the first to share your experience.
Sign in to leave a review for Onyx.
Onyx Alternatives
Looking for alternatives to Onyx? Here are similar tools in the same category:
Learn to Master Onyx
With Skill Leap courses
Get in-depth training on Onyx and 50+ other AI tools with our expert-led video courses.
Start 7-Day Free Trial