Onyx

Name: Onyx
Brand: Onyx

AI Native

Helping modern companies research, create, and automate with LLMs - powered by team specific knowledge.

Est 2023|Raised $10.5M|12 Ref Customers

LinkedIn 4K|Website|Web traffic 232K|FTEs 2|HQ San Francisco, California, US

MarketsKnowledge Management AI Assistant

Published Oct 15, 2026

01Common complaints

- Company sizes or stages where this product is overkill or insufficient - Very small teams (under ~25–50 employees) without a cross-system knowledge problem may find deployment and tuning overhead outweighs benefits. - Industries or use cases where competitors are clearly preferred - If you need the highest immediate relevance on a messy corpus with minimal tuning and want a single accountable vendor with mature SLAs, premium incumbents (for example, Glean) may be a better fit. - Technical requirements that this product cannot meet - If specific identity standards (for example, OIDC or SCIM) are required in a particular edition and are not available, another vendor may be necessary. - Budget constraints that make alternatives more appropriate - If the requirement is for minimal ops and a free solution, a basic open-source chat UI with a single document store may suffice. - Conversely, if the organization prefers a fully managed enterprise solution to avoid operations overhead, a proprietary SaaS vendor may be preferable. - Specific workflows or integrations where the product falls short - If key systems are not covered by Onyx connectors and the organization cannot allocate engineering time to build and maintain custom connectors, an alternative is advised. - Red flags in the buying process - If ROI claims are presented without methodology (for example, a reported 30x ROI), insist on a pilot with measurable KPIs and a documented methodology before committing. - Current limitations acknowledged by the vendor - Not comprehensively enumerated in the sources captured; buyers should request an explicit list of known limitations from the vendor. - Features on the roadmap (not yet available) - note estimated dates if available - Roadmap details and dates were not reliably captured in the reviewed sources. - Scale limitations (at what volume/size does it break down?) - Not publicly quantified in the sources reviewed. Buyers should validate maximum indexed documents, re-index times, and query latency targets under load. - Security or compliance gaps - SOC 2 Type II is claimed on marketing pages, but public audit artifacts were not located. ISO certifications and detailed compliance mappings are not clearly documented publicly. - Technical debt or architectural limitations mentioned in reviews - Public reviews are limited; community discussion points more to feature gating and product maturity concerns than specific architectural debt. - API limitations or integration constraints - Buyers should validate API rate limits, connector sync semantics, and end-to-end enforcement of permissions.

02Why they buy

- Competitive advantages validated by customer feedback (not just vendor claims) - Community feedback highlights: - A combined offering of a self-hostable chat UI, RAG, and connectors. - Broad support for multiple LLM providers. - Unique features and capabilities competitors lack - The most defensible differentiator is the combination of open-source licensing and self-hosting options with an enterprise-search orientation. - Differentiation from competitors based on reviews - Public review-platform validation is limited; differentiation is principally expressed in open-source community sentiment and forum discussions. - Value propositions with evidence from case studies - Vendor case studies report productivity benefits (for example, Ramp); these are vendor-published and should be validated in procurement. - Areas where competitors have advantages (balanced view) - Glean: commonly perceived to provide stronger enterprise-level relevance and connector maturity. - Coveo: established enterprise search/relevance capabilities with extensive adoption in larger deployments.

03Field reports

Important: “Onyx” is a heavily overloaded name online (macOS utility, crypto wallets, unrelated products). The quotes below reference specifically Onyx (formerly Danswer), the open-source enterprise chat/search product. - Gartner Peer Insights - Onyx has a Peer Insights listing, but it currently shows no reviews in the profile snapshot. This means Gartner Peer Insights does not yet provide validated user sentiment for Onyx; buyers should rely more heavily on community discussions and reference calls. - Reddit (community/social) - Positive/announcement context - A project announcement post describes the product as combining a chat UI, RAG, and extended research features; the post emphasizes self-hosting and model flexibility. - Production/self-hosting realism (tempered) - In a data engineering thread about production self-hosting, a commenter described search relevance as adequate for the price but noted it did not match premium incumbents; the comment suggested users should expect lower levels of enterprise polish compared with those vendors. - Security/auth friction (negative/limitation) - In a self-hosted thread, a user commented that the lack of enterprise identity support in community/non-enterprise instances was a blocker, suggesting some identity features may be gated or not fully available in community/self-host editions. - Hacker News (community/social) - In the Launch HN discussion, a commenter questioned the lock-in narrative and suggested lock-in concerns may be less important unless the workflow is strategic. The commenter also described proxy-based approaches to LLM privacy, offering a substantive critique about whether open-source/self-host differentiation is always essential. - G2 / TrustRadius / Capterra - No clearly accessible product listing for Onyx (formerly Danswer) was found in the searches performed. Search results frequently surfaced unrelated products. Practical implication: review-platform sentiment is currently sparse; buyers should prioritize pilots, reference customers, and community validation. - Why customers cite choosing this product over alternatives (from community and vendor statements) - Preference for open source and the ability to self-host. - Wide connector breadth (often cited as many sources). - Desire for a single UI that supports multiple LLMs and integrates web search and extended research workflows. - Why some customers leave for competitors (churn reasons) - Direct "we switched away from Onyx" stories were not prominent in accessible sources. - Community commentary suggests likely churn vectors: - Search relevance that does not match leading premium incumbents for complex/messy enterprise corpora. - Enterprise identity and governance requirements (SSO/SCIM and similar) depending on edition. - Need for more mature enterprise support and product polish.

Case Studies

- Detailed customer success stories with metrics - Onyx publishes a Ramp case study and reports outcomes such as a very high vendor-published ROI and handling a large volume of questions per week. - Specific results and quantified business impact - The high-ROI and volume metrics for Ramp are vendor-published case study figures. The methodology for these metrics is not publicly disclosed in the sources reviewed; treat them as indicative rather than independently validated. - Implementation details and time to value - Publicly accessible details are limited. Buyers should request specifics such as time spent on connector setup, number of sources indexed, and governance steps for permissions and sensitive data. - Challenges faced during adoption and how they were resolved - Not clearly detailed in the accessible excerpts. Buyers should ask reference customers about on-premises deployment, connector edge cases, and governance hurdles. - CRITICAL: Note when metrics are vendor-published without independent verification - The Ramp case study metrics are vendor-published and the methodology is not disclosed; use those figures as directional evidence only.

04Best fit

- Company sizes: specific employee counts and revenue ranges - Likely best fit: - Mid-market to enterprise: mid-market and large enterprise organizations. - Revenue ranges commonly associated with enterprise search purchases: organizations with mid-market to enterprise-level revenues (high variance). - Smaller teams may use Onyx, particularly via self-hosted/community editions, but may find operational overhead proportionally higher. - Industries and verticals where adoption is strongest (with customer examples) - Knowledge-intensive and SaaS-heavy organizations, including engineering/product/support organizations. - Vendor-published customer example: Ramp (case study). - Roles and personas: decision makers AND end users - Decision makers: CIO/CTO, Head of IT, Security leadership, Platform engineering leads. - End users: Support, Sales, Engineering, Product Ops, and general knowledge workers. - Pain points and challenges the product addresses - Fragmented knowledge across Slack/Drive/Confluence/Jira/GitHub/CRM. - Onboarding friction for new employees. - Repeated internal questions and undocumented tribal knowledge. - Qualification criteria: what makes a GOOD fit - Willingness to invest in connectors, permissions, and retrieval tuning. - Requirement for self-hosting or open-source transparency. - Strong internal platform/IT capacity. - Deal disqualifiers: what makes a POOR fit - Limited tolerance for self-host operations and limited IT bandwidth. - Requirement for fully packaged enterprise IAM features in a free/community edition. - Requirement for an established presence on major review platforms, which is currently limited. - Real customer examples with company names when available - Ramp (vendor-published case study).

05The pitch

What the product does (core value proposition in objective terms)

- Onyx is an open-source enterprise AI assistant and search platform designed to make internal company knowledge (documents, tickets, chats, wikis, repos, CRM records) searchable via a conversational interface ("AI chat") and traditional search. It typically uses retrieval-augmented generation (RAG) combined with connectors into common business systems. The product is also described by the vendor as an "application layer for LLMs," supporting multiple LLM models and providing a workspace where teams can create assistants/agents and automate workflows. - Onyx was formerly known as Danswer (rebrand) and is commonly described as a self-hostable chat-style UI for teams with RAG, connectors, and extended research features.

Target market and market positioning

Primary market: mid-market to enterprise teams that want AI chat/search over internal knowledge while retaining control over data (self-hosting option) and avoiding vendor lock-in.
Positioning: an open-source enterprise search and AI assistant alternative that competes with proprietary enterprise search/AI assistant vendors and other self-hostable/open-source chat + RAG projects.

Get a comprehensive analysis of Onyx including market position, competitive landscape, adoption trends, and peer benchmarks.

Download full report →