Top 5 Onyx (Danswer) Alternatives (2026)

Onyx, formerly Danswer, is open-source AI chat and enterprise search over your company’s apps. Its Community Edition is MIT-licensed and self-hostable, and Onyx Cloud Business costs $20 per user per month billed annually. Buyers who look for an Onyx alternative usually hit one of three facts from Onyx’s own pages:

• Self-hosting is a stack to run. The full Standard deployment is a search index, background workers, AI model servers, Redis and MinIO. Onyx’s sizing guide asks for at least 4 vCPU and 10 GB of RAM and warns that the index blocks all writes when the disk reaches 95%. Someone on your team patches, upgrades and watches it for as long as you use it.
• The managed cloud defaults to the US. Onyx Cloud is hosted on AWS with US-east as its default region, and lists OpenAI and Anthropic as default language-model subprocessors.
• Teams is not there yet. Onyx’s integrations page offers answers in Slack and lists Microsoft Teams as “coming soon”. Its Teams connector indexes channel conversations, channel files and meeting transcripts and chats.

Here are the five alternatives worth shortlisting in 2026. For the direct head-to-head, see amaiko vs Onyx.

1. amaiko: data in Germany, nothing to operate, right in Teams

amaiko keeps your company data under European law without a stack to run. It is a managed service hosted in Germany on amaiko’s own infrastructure, with AI processing in the EU by default, GDPR-compliant and aligned with ISO 42001, the management standard for AI systems. There is no cluster to deploy and no index to size: amaiko handles updates, security and scaling.

It works in Microsoft Teams. Your team installs amaiko in Teams and starts typing in the chat that is already open. There is no separate portal and no training programme. amaiko reads and drafts across email, calendar, files, Word, Excel and PowerPoint, SharePoint and OneNote, and each person signs in with their own account.

It acts before you ask. Your morning briefing is ready when you open Teams, and once you switch on Active Inbox, it sorts your email before you read it.

It remembers. amaiko builds a persistent corporate memory of projects, customers, decisions and how each person works. Overnight it learns from each person’s own conversation with amaiko: personal matters stay personal, and after a privacy check, objective facts such as processes, deadlines and responsibilities are filed for the team or company, so context stays when people leave. For your team that means 35% less time spent searching and onboarding up to 57% faster.

It connects to your business systems through its marketplace: HubSpot, Dynamics 365 Business Central, Personio, Jira, Confluence, DATEV and more.

The price is one number. amaiko costs €41.58 per user per month billed annually (€499 per user per year), or €49.90 monthly, from five seats, with hosting in Germany and unlimited tokens included. The full calculation against Onyx’s list prices is in the comparison.

Best for: companies that work in Microsoft Teams and want their data in Germany without staffing a self-hosted platform. Book a demo to see it in your own Teams.

2. Glean: enterprise search, price on request

Glean is an enterprise search and AI agent platform. Glean does not publish prices: its pricing link leads to the home page, whose customer stories are large enterprises such as Booking.com with 14,000 employees. For a mid-sized company, the budget conversation starts with a sales call. Glean also offers a Customer Hosted model that it deploys inside your own GCP or AWS environment, which your team prepares before setup.

Full comparison: amaiko vs Glean.

3. Dust: agents you build, billed in credits

Dust is a platform for building AI agents on company data. Its Business plan is for teams of up to 100 people: the Pro seat costs €24 per month billed yearly (€30 monthly) with 8,000 credits per seat per month, and the Max seat €120 billed yearly with 40,000 credits. Dust offers US and EU data residency; single-tenant deployment is part of its Enterprise plan. In Teams, Dust runs as a bot that an administrator uploads as a custom app. For your team that means agents someone has to design, and a credit allowance to watch.

Full comparison: amaiko vs Dust.

4. meinGPT: German hosting, model costs on top

meinGPT is an AI platform from Germany that says it runs on Hetzner infrastructure in German data centres. Its price list charges €18 per user per month for the first 50 users, including a 10% annual discount. The licence excludes model costs, which meinGPT plans at €2.31 per user per month by default, and Microsoft Entra ID group sync and SCIM provisioning sit in a governance add-on at €1.35 per user per month. Connecting sources inside your own network takes its Outpost add-on, from €225 per month. For your team that means a licence price plus a usage budget to manage.

Full comparison: amaiko vs meinGPT.

5. Langdock: an AI workbench with a Teams side door

Langdock is a chat, agent and workflow platform. Its Business plan costs €23.20 per user per month for a Business seat and €99 for a Business Max seat, billed annually (Business seats: volume discounts above 50) and excluding VAT; workflows beyond the starter package are a paid add-on. By its security page, the standard service runs on Microsoft Azure in the EU, a dedicated deployment starts at 2,000 seats, and your own cloud or on-premise at 5,000. Langdock has a Teams app, which its own changelog describes for “small ad-hoc situations” such as quick translations “without needing to open Langdock”. For your team that means the main work still happens in Langdock.

Full comparison: amaiko vs Langdock.

How to decide

Teams leaving Onyx decide on five things. Here is where the alternatives stand, and where amaiko lands.

Run it yourself, or have it run. On Kubernetes, the default container requests of self-hosted Onyx add up to about 12 CPU and 24 GB of memory. Your administrators also connect the language models and keep a Microsoft Entra app alive for the Microsoft 365 connectors: when its secret or certificate expires, indexing stops. Of the alternatives, Glean can run inside your own cloud, Dust sells single-tenant deployment in Enterprise, and Langdock goes on-premise from 5,000 seats. amaiko has nothing for you to operate, and no on-premise version.

Where the data lives. Onyx Cloud defaults to AWS in US-east; region-specific deployments are on its Enterprise list, with “Contact us” instead of a price. Dust offers US and EU residency, meinGPT says it runs in German data centres, and Langdock’s standard service runs on Azure in the EU. amaiko stores your data in Germany on its own infrastructure, each customer in its own isolated environment with its own databases. AI processing runs in the EU by default; data leaves the EU only if you explicitly book a model that runs elsewhere.

Search over documents, or an assistant that works in Microsoft 365. Onyx syncs your sources into its own index. Its Microsoft 365 connectors sign in as an application, not as a user, with permissions only a Global Administrator or Privileged Role Administrator can grant, and by default the app can read every SharePoint site in the tenant. amaiko works in Teams, signed in as each person with that person’s permissions. It searches Teams chats, channels and documents on request, drafts emails that go out after you approve them, writes Word documents, schedules meetings and posts Teams messages.

Connector coverage. Onyx’s pricing page lists 40+ app connectors. Connectors that mirror source permissions are an Enterprise Edition feature, and Onyx warns that a public connector reading private data shows it to every Onyx user. amaiko covers Microsoft 365 end to end, plus Google Workspace and Box, and its marketplace specialists work inside business systems from HubSpot to DATEV.

Total cost in year one. amaiko costs €4,990 for 10 people and €12,475 for 25, billed annually, with hosting, operation and unlimited tokens included. Onyx Cloud Business costs $20 per user per month billed annually, in a cloud that defaults to the US; Enterprise has no published price. Self-hosting the Community Edition costs nothing in licences; by Onyx’s docs, you provide the servers, the language models, the Entra app and someone to run upgrades. Onyx puts no figure on that part.

If your company works in Microsoft Teams and you want your data in Germany without running a platform, start with amaiko. Book a demo, or compare the numbers in detail in amaiko vs Onyx.