Snapshot
Continue to About ↓- What it does
- TensorX is a managed AI inference provider that runs open-weight language and embedding models on its own GPUs in EU data centres and sells access through an OpenAI-compatible API billed per token.
More detail
- It is built around zero data retention (prompts and completions held in volatile memory only, not logged or used for training) and EU-only data residency.
- Best for
- EU enterprises in regulated industries such as healthcare, finance and legal that need to run LLM inference under GDPR and EU AI Act expectations without data leaving EU jurisdiction.
More detail
- Also a fit for developer teams building agents, RAG pipelines, conversational AI or coding assistants who want a one-line, OpenAI-compatible migration to open models.
- Pricing
- Developer tier is pay-as-you-go per token with prepaid credits, no monthly fee and no stated minimum
More detail
- input and output prices are listed per 1M tokens (for example DeepSeek V4.1 Flash at $0.50 input and $1.50 output, GLM-5.3 at $1.75/$4.50, Kimi K3 at $3.00/$15.00, Qwen3 embedding at $0.02 input), w
- Main limitation
- Buyers who need certified ISO 27001, SOC 2 Type II or ISO 42001 today, since the first is only 'in progress' and the latter two are planned.
More detail
- Also poor fits are teams that require frontier proprietary models such as GPT or Claude, or that need compute or data residency outside the EU.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- TensorX is a managed AI inference provider that runs open-weight language and embedding models on its own GPUs in EU data centres and sells access through an OpenAI-compatible API billed per token.
- It is built around zero data retention (prompts and completions held in volatile memory only, not logged or used for training) and EU-only data residency.
What it helps with
- Keeping prompts and completions off foreign-hosted infrastructure while still using modern LLMs, via zero data retention
- Handling GDPR controller/processor mechanics for AI workloads, with a DPA, SCCs, Article 33 breach notification and EU A
- Reducing migration friction from OpenAI SDKs by changing a base URL rather than rewriting integrations.
- Lowering inference cost by using open-weight models instead of frontier proprietary models.
- Getting dedicated capacity without running GPU operations, via a privately tuned endpoint.
- Giving regulated buyers an inference layer whose data handling can be described to compliance reviewers.
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| OpenAI-compatible API | Yes | POST /v1/chat/completions with cURL, Python and Node examples; swap the base URL to migrate. |
| Zero data retention | Yes | Prompts and completions held in volatile memory only, not logged, stored, shared or used for training. |
| EU-only infrastructure | Yes | GPU servers in Dublin and Helsinki, with Paris listed on the dedicated tier; data stays in EU jurisdiction. |
| Model catalog depth | Limited | Pricing page lists 13 pay-as-you-go models (DeepSeek, Z.ai GLM, Moonshot Kimi, MiniMax, Qwen); marketing claims of 30+ and 42+ models are unverified. |
| Frontier proprietary models (GPT/Claude) | No | Catalog is open-weight and licensed model families; no OpenAI or Anthropic models. |
| ISO 27001 certification | Limited | Status is 'in progress (infrastructure partner certified)'; marketing says 'ISO 27001-ready'. |
| SOC 2 Type II / ISO 42001 | No | Both listed as planned on the trust and security page. |
| Dedicated sovereign GPU clusters | Yes | Physically isolated cluster, fixed monthly pricing, fixed SLA; custom quote via sales. |
| Prompt caching | Yes | Cache-read line item on LLM SKUs at roughly 25% of the input rate. |
| Alerting to low balance | Limited | Prepaid credits only; API keys auto-disable when balance falls below $0.05. |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- tensorx.ai
Technical profile details
Record ID
9ad68b4f-c325-4146-988f-d68d888f0b7dCanonical URLhttps://profiles.scribble.network/listing/tensorx-9ad68b4f