← Back to Directory
Logo: Myra AI Workspace

Myra AI Workspace

German AI workspace combining chat, projects, agents and an AI gateway with PII masking and Myra-hosted EU models on BSI-attested infrastructure.

Partially EU hostedSustainability claimed
ai-workspaceai-gatewaypii-maskingeu-hosted-modelsbsi-c5germanygdprssoscimmcp
Visit WebsiteFeedback & corrections

Myra AI Workspace is a Munich-built AI work environment from Myra Security GmbH that combines an employee-facing chat and project UI with a multi-tenant AI gateway. Organisations route requests to more than twenty external model providers or to Myra's own EU-hosted open-weight fleet, under central budgets, guardrails and audit logging.

It exists for companies and public bodies that want ChatGPT-class productivity without leaving model choice, residency and personal-data handling to unmanaged consumer accounts. Colour-coded chat modes show whether a turn stays on Myra models, uses frontier models with automatic PII masking, or sends unmasked content where admins allow it.

The concrete differentiator is the pairing of adoption tooling (projects, agents, workflows with human-approval steps, MCP connectors) with gateway controls on infrastructure Myra operates in the EU, from a German KRITIS-oriented security vendor that also publishes ISO 27001 (IT-Grundschutz) and BSI C5 Type 2 credentials for its platform.

EU-operatedAI gateway + workspacePII maskingBSI C5 Type 2 (claimed)SSO + SCIMMyra EU model fleet

Shortlist when you need a German-operated AI chat and gateway stack with Myra-hosted EU models, PII masking and IdP/SCIM controls for regulated teams. Skip when you only want a single-vendor Copilot inside Microsoft 365, or when every workload must be fully air-gapped on your own GPUs without a SaaS control plane. Consider Nextcloud plus a self-hosted model stack, or Microsoft 365 Copilot, depending on whether collaboration files or Office integration dominate.

Key capabilities

Employees work in a familiar chat and shared project workspace with model switching. Green marks Myra-hosted EU models, orange marks frontier models with PII masking, and red marks unmasked frontier use where administrators permit it.

Before prompts reach external providers, configurable guardrails can detect and replace personal data (names, emails, IBANs and similar) and block jailbreak or injection patterns. Tier 1 runs in-process; tier 2 optional sidecars add NLP PII and prompt-guard checks. Masking is skipped on wholly Myra-local legs because data does not leave Myra.

Open-weight models such as Qwen (docs cite the current default Myra on-prem chat model) run on Myra's EU fleet without a BYOK key. Gateways can enable EU-region routing that returns HTTP 403 data_residency_blocked for non-EU routes, including post-dispatch host re-checks when the Myra router would overflow off-EU.

A unified compatibility endpoint resolves twenty-plus providers (OpenAI, Anthropic, Gemini/Vertex, Azure OpenAI, AWS Bedrock, Mistral and others). Admins store provider keys in an AES-256 vault, set fallbacks, exact-match response caching, and organisation/team/user budgets with hard enforcement.

Documented agents support versioning, scheduling, webhooks and required approval before tool runs. Workflows include human-approval steps. MCP connectors attach external tools and data sources; the connector catalog documents Microsoft SharePoint, OneDrive and Graph with Entra OAuth.

SSO authenticates existing users via OIDC or SAML 2.0 (including Microsoft Entra ID) with domain allowlists. SCIM 2.0 on the admin API provisions and deactivates users and groups. Access can be limited by IP allowlist (IPv4).

At a glance

HQ
Munich, Germany
Legal entity
Myra Security GmbH (HRB 202428)
Product URL
ai.myra.eu
Founded (company)
2012
Deployment
Hosted SaaS (Myra EU infra)
Identity
OIDC, SAML 2.0, SCIM 2.0

Best fit when

  • Public sector, finance, insurance or other regulated EU orgs that need central AI governance
  • Teams that want both an employee chat/projects UI and an OpenAI-compatible gateway in one contract
  • Workloads that can stay on Myra-hosted EU open-weight models for sensitive content
  • IT teams rolling out Entra ID / OIDC / SAML SSO with SCIM provisioning
  • Buyers already evaluating Myra Security for KRITIS-oriented edge security and wanting the same vendor for AI adoption

Poor fit when

  • Organisations that mandate fully on-prem, air-gapped LLM hosting with no SaaS control plane
  • Teams standardised on Microsoft 365 Copilot or Google Gemini only and unwilling to add another workspace
  • Use cases that require continuous frontier-model quality with a hard ban on any US provider path (possible only if you lock to Myra-local models and EU search)
  • Buyers seeking an open-source, self-hosted chat UI as the primary deliverable

Consider instead when

  • When: You mainly need Office-integrated Copilot inside an existing Microsoft estate

    Consider: Microsoft 365 Copilot

    Deeper Word/Excel/Outlook integration; single-vendor jurisdiction and model stack.

  • When: You want EU file collaboration with optional self-hosted AI rather than a managed AI gateway

    Consider: Nextcloud

    Strong EU collaboration suite; AI features are not a Myra-class multi-provider gateway.

  • When: You only need European IaaS/GPU cloud, not an AI workspace product

    Consider: STACKIT or Scaleway

    Infrastructure layer; you still build the adoption UI and guardrails yourself.

Jurisdiction & ownership

Legal entity
Myra Security GmbH
US parent / control
No known US parent
CLOUD Act exposure (indicative)
Medium
Hosting / residency
Domain / serviceRoleOperatorStatusEvidence
ai.myra.eu / Myra EU fleetCore app + local model inferenceMyra Security GmbH (EU)EU hostedProduct FAQ, docs Myra EU fleet and EU residency pages, AI privacy policy (2026-10)
Frontier LLM providers (OpenAI, Anthropic, Google, Azure, Bedrock, …)Optional model inferenceMultiple (mostly US) via subprocessor registerNon-EUai.myra.eu/privacy §5.1; product provider list
Web search (Brave or Linkup)Optional search toolBrave Software, Inc. (US) or Linkup (France)Partialai.myra.eu/privacy §5.2

No known US parent. CLOUD Act risk is path-dependent: Myra-local models can avoid US providers; enabling frontier models or Brave search reintroduces US-group processing. Indicative only, not legal advice.

Sustainability

Vendor claimed

Company certifications page states renewable energy for defense infrastructure, ISO 14001 and ISO 50001, and ClimatePartner compensation. Not independently verified here.

  • Independent security / no-logs auditNot found
  • ISO 27001Vendor claimed
  • SOC 2 / SOC 3Not found
  • BSI C5 Type 2Vendor claimed
  • +4

Considerations & known limitations

  • MediumOptional US frontier model path

    Out of the box the product markets 20+ providers including major US labs. Without EU-region lock and provider allowlists, prompts can leave the EU even though the control plane is German-operated.

  • LowCompany certs vs product scope

    ISO 27001 IT-Grundschutz, C5 and related seals are published for Myra's security platform. Confirm in writing that AI Workspace production is in scope of the current attestation period.

  • LowMyra fleet overflow edge case

    Docs describe that a saturated Myra EU fleet router can overflow to a cloud fallback; residency-armed gateways re-check the observed host and fail closed, but residency routing is off by default.

  • LowLive subprocessor register requires Trust Center

    The authoritative provider list is the live register on the Trust Center / privacy page. Review the current German-language entries before contracting; do not rely on a static blog list.

Open questions for due diligence

  • Will procurement receive the current BSI C5 Type 2 and ISO 27001 packs with AI Workspace explicitly in scope?
  • What is the default provider allowlist for a new production tenant (Myra-only vs frontier enabled)?
  • Can Linkup and Myra-local models fully replace Brave and US frontier providers for your residency policy?
  • Is a signed AVV/DPA for AI Workspace available without NDA, and does it list the live subprocessor register?
Request vendor DPA / enterprise contact
Visit WebsiteCompare similar toolsView alternativesFeedback & corrections

Frequently Asked Questions

Always through Myra's EU-operated control plane first. If the selected model is on the Myra EU fleet, inference can stay there. If admins enable external providers (OpenAI, Anthropic, Google, Azure, Bedrock and others), prompts are forwarded to those providers under the product subprocessor register and, for BYOK connections, under your own provider contract. Optional EU-region routing fails closed for non-EU routes. Web search defaults to Brave (US) unless EU residency enforcement selects Linkup (France).

Yes. Admins control which providers and models a gateway may use, can disable individual Myra models per gateway, and can arm EU-region routing so non-EU upstreams are rejected. Chat UI colour coding shows users when they are on Myra-local versus masked or unmasked frontier modes.

Provision users manually or via SCIM 2.0, then enable SSO with OIDC and/or SAML 2.0 (Entra ID has a dedicated guide). SSO signs in existing accounts only; it does not auto-create users. Allowed email domains bound the IdP to the tenant.

Seat-based licensing plus a platform choice (EU-located versus EU-governed Myra Cloud with local LLMs) and model usage. Token costs can be billed through Myra or kept on your own provider contracts via Bring Your Own Key. Exact list prices are not reproduced here; confirm current packaging on the official site or with sales.

No. Myra CDN is the company's Anycast CDN and edge security platform. AI Workspace is a separate AI adoption and gateway product on ai.myra.eu, sharing the same Munich legal entity and overlapping infrastructure certifications.