amberSearch Pricing 2026: What amber Really Costs Today

amber (formerly amberSearch) costs €12.99 per user per month in Starter, €17.99 per licence per month in Business (from 50 licences), and Enterprise is priced individually from 200 users. Starter can be booked from a single user and, according to the vendor, includes chat with premium models, agents that execute more than 1,000 actions in over 120 systems, and connections to any number of systems via MCP. An optional Data Layer extension connects up to 10 systems through deep integrations. AI search together with a Data Layer of up to 15 connectors is included in Business.

amber is a German company without US shareholders and focuses on the Mittelstand. This page explains what each tier contains, what you actually pay per month at different team sizes, and how the credit model for premium models works. All prices exclude VAT and come from the vendor pricing page (amber.de/preise), as of 14 October 2026. Pricing pages change, so check the current version before you sign.

amber sells three tiers: Starter per user, Business per licence with a 50-licence minimum, and Enterprise on request from 200 users. The per-head price is only half of the decision. The other half is whether you need AI search over your company knowledge as a fixed part of the package, which is what Business adds, or whether chat, agents and connections via MCP are enough, which Starter already covers.

The vendor's own calculator makes the difference concrete: 50 users cost €649.50 per month on Starter and €899.50 on Business. Starter has no minimum, so a team of ten pays for ten seats. If you are evaluating amber as an AI context layer for your existing tools, decide on the search and connector question first, then multiply.

Business is more than Starter plus search. It adds its own environment and database, SSO via Entra ID (Azure AD) and a dedicated customer success manager. Enterprise adds more than 15 connectors, bring-your-own LLM key, a separate staging environment, individual SLAs and an individual domain, and is aimed at large organisations with their own infrastructure and compliance requirements.

All tiers can be tested free for 30 days. Premium model usage runs on a credit model across all tiers, explained further down.

Starter is the entry point for single users and small teams, and it is a full product, not a demo tier. According to the vendor it includes chat and agents, more than 1,000 actions in over 120 systems, plugins and a desktop app, API access, always-on models without a limit and premium models from OpenAI, Anthropic, Google and Mistral through a daily allowance per user. Systems can be connected via MCP; the optional Data Layer extension adds deep integrations for up to 10 systems.

What Business adds on top is AI search with a Data Layer of up to 15 connectors as part of the package, plus its own environment. If your main goal is to break down data silos with one search across many sources, compare Starter with the Data Layer extension against Business for your headcount.

On Starter you pay exactly for the people who use it: 10 users cost €129.90 per month, 50 users €649.50. Business starts at 50 licences, so its entry point is €899.50 per month (€10,794 per year). Above 50 people the two tiers differ by €5 per head, which is the price of AI search, the Data Layer with up to 15 connectors and your own environment.

Below 50 people the question is different: Starter is the tier designed for that size, and the Data Layer extension is the way to deep integrations without the Business minimum. Premium model usage beyond the daily allowance comes on top as credits, the same in every tier. That is the second number to budget for, and it depends on usage, not on seats.

amber no longer sells fixed token packages: premium models run on a daily allowance per user and, beyond it, on company-wide credits. Companies define the credit amount freely, either as a one-off top-up or as a recurring monthly top-up. Always-on models remain usable without limit after the daily allowance is used up.

API usage is not included in the seat price and is billed in credits. According to the pricing page, the listed model prices include a 5 % surcharge and the models run in EU regions. For budgeting, that means one predictable number (seats) and one usage-driven number (credits) that you can cap yourself.

amber targets the Mittelstand and can be booked from a single user, so the price list no longer excludes small teams. Founded in Aachen in 2021 by RWTH graduates, the company raised a €2.1 million seed round in its early days, as reported by the Aachener Zeitung, and today states more than 400 business customers. It describes itself as a German company without US shareholders.

According to the vendor, all development and processing steps take place on servers in Germany and Europe, with ISO 27001 certification, GDPR and EU AI Act compliance. That is a serious data sovereignty story. Whatever tool you pick, check how it handles your permission architecture, because a search across systems is only as safe as the access rights it respects.

The useful comparison is not price per seat but where the product starts: from search, from chat, or from team context. US platforms like Glean publish no list prices and are sold as enterprise contracts. amber starts from AI search plus chat and agents. It is worth understanding how enterprise search differs from a context layer before you compare two price lists, because the words are used loosely and the products are not the same.

Our own product, Teamo AI, starts from team context. It connects an unlimited number of systems: Slack, Teams, Jira, Notion, HubSpot, Pipedrive, your calendar and any in-house API through the AI integration assistant. Agents work across those systems, and you choose between models from OpenAI, Anthropic and Google or European models from Mistral, all EU-hosted, from the first seat at €9.97 per user, with no minimum. Teamo AI works where your team already talks. Which fits depends on whether your problem is finding documents or coordinating a team. Our comparison of European ChatGPT alternatives and the AI strategy guide for the Mittelstand cover the decision, and the vendor lock-in risks are worth checking before any signature.