Kisuno AIPricing Docs BlogENSelf-host for free Get startedReady for production. Request your access.The European AI platform for your company: GPT, Claude, GLM, Kimi and DeepSeek through one account, your data in Switzerland or on your own server.// Free to get startedCHF 50 (EUR 55) of model creditTesting is free and not limited in time. No payment details are required.

You confirm your address via the link in the email.

We look at your request and set up your access.

You start with the full feature set, with CHF 50 (EUR 55) of model credit ready to use.

Because demand is high, we release access in stages. We get back to you once we have looked at your request.Work email address *Company optionalTeam size optionalPlease choose …12–1011–5051–200200+Interested in optionalHosted in SwitzerlandSelf-hostingBothRequest accessYour details stay with us and are not shared with anyone. We only write about your access and the launch, and you can unsubscribe at any time. Privacy policyPrefer to see it first? Book a live demoPrefer to try it yourself? To the installation guideFrequently asked questions.What prospective customers ask us before signing up, and what we answer.Open all answersHow signing up worksDoes the product already exist, or is this just a page with a form? +The platform is ready for production use and in internal use at Herr Informatik GmbH. The complete operating documentation is openly available: system requirements, installation, configuration and commands. What is still pending today is not the software but the order in which we open new accounts.View the documentationWhat happens after I have submitted the form? +You receive an email with a confirmation link, valid for 48 hours. Once you have confirmed, the request reaches us and is processed step by step.How long does it take until I have access? +Because demand is high, we set up accounts in stages, one request at a time. Anyone who wants to see the platform straight away books a live demo.Book a live demoWhy do you ask for a work email address? +Every account is assigned to an organisation, so the form does not accept addresses from public mailbox providers. Companies without a domain of their own are sorted out in a call.Book a live demoWill I get advertising or phone calls afterwards? +No. Your details stay with us and are not shared with anyone. We do not ask for a phone number.Am I committing to anything if I take part? +No. There is no obligation to give feedback, no meetings and nothing to sign. You can stop at any time.Can I not simply install the software myself instead of asking? +Not yet: the GitHub repository and its publication under the MIT licence are coming soon. The complete installation guide is already openly available. Until then, managed access is the way to work with the platform.To the installation guideStarting credit and costsWhat exactly is the starting credit of CHF 50 (EUR 55)? +A one-off credit for the model usage on your account. Every request to a model costs tokens, and those tokens are charged against the credit. For as long as it lasts, no costs arise.How far do I get with CHF 50 (EUR 55)? +That depends on the choice of model and on how you use it: short text work is cheap, analysing long documents with the largest models considerably more expensive. So we do not quote a fixed number of requests. Once the credit is used up, usage stops, no invoice arises, and you decide whether to carry on.Do I have to provide payment details? +No, neither for the request nor for the access itself. There is no card on file that would be charged when the credit expires.What happens when the credit is used up? +Usage stops, and no invoice arises. You then decide whether managed operation should continue: CHF 7 (EUR 7.50) per person per month. Through the end of the current month: 50 percent off with a twelve-month term, that is CHF 3.50 (EUR 3.75) per person per month.View all pricesHow do we keep costs under control when billing follows usage? +Usage is broken down per person, group and model, and a shared monthly budget caps it for the whole organisation. Once it is reached, a soft block takes effect: requests are stopped rather than quietly running on, and the administration can raise the budget or release individual people. An additional monthly limit per person can be set.Is there a minimum number of users? +No. Self-hosting is free from a single person upwards and has no seat limit, managed operation is billed per person. There is neither a package for a fixed company size nor a setup fee.Does the starting credit also apply if we host it ourselves? +No. It sits on a managed account, because only there does the model usage run through us; with self-hosting you settle directly with your model provider. The form accepts both at once, and a later move from managed operation to your own server is provided for.Data and confidentialityAre our inputs used to train models? +No. Customer data does not train any models, neither at the inference provider nor at our end. Telemetry is switched off by default in all components, and no tracking takes place without active consent.You offer the large models from OpenAI and Anthropic. Does that mean our data goes to the United States after all? +Conversations, files and accounts are held in Switzerland. Kisuno offers the current models from OpenAI (ChatGPT) and Anthropic (Claude); they run in the AWS data centre in Frankfurt. AWS is a US company, so the US CLOUD Act applies even with servers in Frankfurt, and we say so openly. For customers with higher data-protection requirements, open frontier models such as GLM, Kimi and DeepSeek are available, operated by European companies on servers in Europe; the CLOUD Act does not apply there. Whatever model you choose: inputs and uploaded data are neither stored by the model provider nor used for training. Two exceptions we state openly. First, AWS automated abuse detection: it can flag messages, for instance on signs of attacks, malicious code or prohibited content. AWS retains flagged messages with the GPT models, and all requests with Claude Fable 5 and 5.1, for up to 30 days to review and stop misuse; none of it is shared with the model providers or used for training. Second, web search: it passes the search query on to external search engines without the identity of the user, so we do not claim across the board that everything stays in Europe.The data paths in detailWhat good is Swiss hosting if the US CLOUD Act applies anyway? +Besides the location, what counts is who controls the operator. Behind Kisuno stands Herr Informatik GmbH in Windisch, Switzerland, with no American parent company; operations run in the data centres of Green in Lupfig. With the models it is different: the models from OpenAI and Anthropic run at AWS in Frankfurt, and the US CLOUD Act applies there. Anyone who wants to rule that out picks the European model tier with GLM, Kimi and DeepSeek, or runs the platform on their own hardware.How long do you store our conversations? +Out of the box no retention period is set; conversations stay until users or the administration delete them. A period can be configured freely in the admin panel, after which expired content is removed together with its search index entries. Exporting and deleting the data of an individual person is carried out by the administration there as well.Can anyone at your end read our chats? +Anyone who self-hosts gives no third party any access. In managed operation the operator accesses the system only for maintenance, governed by the data processing agreement and the list of subprocessors. Inside your instance, users see exactly what their role and group allow, and every access to a connected system is logged.Do we get a data processing agreement, and who are your subprocessors? +Yes. The full contract text and the named list of all subprocessors, with purpose and location, are openly available on this website, before signing up and without a sales call.Read the data processing agreementDo you have ISO 27001 or SOC 2? +Certified to ISO 27001 are the data centres of Green in which Kisuno is operated; the certificates are held by Green and not by us. A certification of our own to ISO 27001 is in preparation. The technical and organisational measures are documented openly.To the measuresWe work under professional secrecy. Can we use the tool at all? +That assessment rests with you and your supervisory body. The basis for it is in place: the data processing agreement, the named list of subprocessors, storage in Switzerland, model processing in Europe, on request exclusively at European providers, and operation on your own hardware. This is not legal advice.Product and operationsWe already have ChatGPT or Microsoft Copilot. Why add something else? +ChatGPT in the browser is a personal account with no connection to your company systems and no central control, and Copilot is tied to one environment and one model. Kisuno is a company instance with roles, permissions, an audit log and a cost overview, with the current models from OpenAI (ChatGPT) and Anthropic (Claude) plus open frontier models such as GLM, Kimi and DeepSeek to choose from and with your own systems connected. The models are partly the same; the difference lies in the data path, in the choice and in the control.The comparison in detailWhy not simply run Open WebUI or LibreChat ourselves? +That is possible, and Kisuno is itself a thin fork of LibreChat; self-hosting stays free and without a seat limit. On top of that come permissions down to tool level, an audit log, usage and budget with a soft block, prepared integrations and, on request, operation including updates and backup.Which models do we get, and do we really get the one we chose? +The current models from OpenAI (ChatGPT) and Anthropic (Claude), plus open frontier models such as GLM, Kimi and DeepSeek, operated by European companies on servers in Europe. The administration curates the selection and sets a default model; usage is broken down per model and can therefore be traced.Can we connect our own documents? +Yes, as a knowledge base; SharePoint from Microsoft 365 and ERP systems through the MCP gateway are prepared as well. How good the answers turn out depends on the material: cleanly typeset texts work well, scanned forms and nested tables are difficult for any system. A clearly delimited set of documents gives better results at the start than a whole drive.View all integrationsDo we need an IT department, and how quickly are we ready to start? +Not for managed operation: the operator handles operation, updates and backup, and you manage users, roles and permissions in the admin panel. The technical part takes minutes. What comes afterwards takes longer: defining roles, connecting systems, working through the first use case. A productive rollout takes days rather than minutes.Half of our people use ChatGPT privately today. Will this get it under control? +On a company instance, usage is broken down per person, group and model, every access to a connected system appears in the audit log, and roles determine who sees which models and knowledge sets. How far that visibility reaches is set through roles and permissions; a rule settled in advance makes the rollout easier.Still something open? To the full collection of questionsKisuno AIYour AI coworker. Stays in Europe.Deployable anywhereManaged, own cloud or on-premGDPR-compliantEU deploymentMore about securitySwiss instance · Lupfig --:--:--ProductFeatures Pricing Sovereignty Integrations Comparison Self-Host FAQ Documentation Blog Our vision Tea break Book a live demoTrustSecurity & Trust Governance DPA TOM overview SubprocessorsLegalImprint Privacy Terms (GTC)LanguageDeutschEnglishFrançaisItalianoEspañol© 2026 Kisuno by Herr-Informatik GmbH

{{ t.h1a }}{{ t.h1b }}

{{ t.sub }}

{{ t.credit.eyebrow }}

{{ creditAmount }}{{ t.credit.unit }}

{{ creditBody }}

{{ s }}

{{ t.hint }}

{{ eLink }}

{{ t.labels.email }}*

{{ eEmail }}

{{ t.labels.company }}{{ t.labels.optional }}

{{ t.labels.users }}{{ t.labels.optional }}

{{ t.labels.interest }}{{ t.labels.optional }}

{{ eForm }}

{{ t.labels.note }} {{ t.labels.noteLink }}

{{ t.alt.callLead }} {{ t.alt.callLink }}

{{ t.alt.selfLead }} {{ t.alt.selfLink }}

{{ t.pending.eyebrow }}

{{ t.pending.title }}

{{ pendingBody }}

{{ t.pending.spam }}

{{ t.success.home }}

{{ t.success.eyebrow }}

{{ t.success.title }}

{{ successBody }}

{{ t.success.home }}

{{ t.success.ask }}

{{ t.labels.usecase }}{{ t.labels.optional }}

{{ t.success.askThanks }}

{{ t.faq.h2 }}

{{ t.faq.lead }}

{{ g.label }}

{{ it.q }}

{{ it.a }}

{{ it.linkText }}

{{ it.linkText }}

{{ it.linkText }}

{{ it.linkText }}

{{ it.linkText }}

{{ it.linkText }}

{{ it.linkText }}

{{ it.linkText }}

{{ t.faq.moreLead }} {{ t.faq.moreLink }}