Blog
What to know before you trust an AI with your data.
Sovereignty, open-weight models, GDPR, the CLOUD Act, the AI Act: articles written to help you decide, with every fact sourced and dated.
Sovereignty & compliance
- 13 Oct 2026European inference providers: who actually runs AI models in EuropeWho runs AI models in Europe: OVHcloud, Scaleway, Mistral AI, IONOS, STACKIT, AKI.IO, Infomaniak, Nebius and more. Company, data centres, prompts.
- 20 Oct 2026Is OpenAI GDPR-compliant? What ChatGPT and the API do with your dataOpenAI offers EU contracts and transfer clauses, but it remains a US company within reach of the CLOUD Act. What that means for ChatGPT, the API and you.
- 20 Oct 2026GDPR and the CLOUD Act: why hosting your AI in the EU is not enoughThe GDPR governs transfers; the CLOUD Act follows the company. Why an EU data centre run by a US group stays within reach, and how to check a provider.
- 20 Oct 2026Sovereign AI: build your own cluster, or use an API?Palantir and NVIDIA's on-premises sovereign AI design starts at 16 servers and 32 GPUs. What owning buys, what it leaves open, and when an API is enough.
Models
- 20 Oct 2026Tool calling on open-weight models: why it depends on who runs themTool calling turns a model into an agent, but it depends on the serving stack as much as on the weights. How to check it in one call, and which models pass.
- 20 Oct 2026Open-weight reasoning models: what thinking costs, and how to choose oneReasoning models think before they answer, and that thinking is billed as output. Which open-weight models reason, what it costs, and how to choose one.
- 20 Oct 2026Open-weight AI models in 2026: which to use for what, and who runs themThe open-weight models of 2026 — GLM, Qwen, DeepSeek, GPT-OSS, Llama, Mistral — sorted by use, with size, licence, price and the company that runs them.
- 20 Oct 2026What is open-weight AI, and what can you do with the weights?An open-weight model publishes its weights, the numbers it learned in training. What you get, what you don't, how much memory they need, and who can run them.
- 6 Oct 2026What is open-source AI?Open-source AI lets anyone use, study, modify and share a model, with its weights, code and training-data information. What qualifies, and what it changes.
- 20 Oct 2026French open-source AI: the companies, the models, and where they runMistral AI, Hugging Face, OVHcloud, Scaleway: who does what in French open-source AI, and why a model's origin does not decide where your data goes.
- 13 Oct 2026DeepSeek V4.1 Flash: what it is, and where it runsDeepSeek V4.1 Flash: a 552B multimodal MoE with a 1M context, under MIT, released on 10 September 2026. On Evanesca, AKI.IO runs it in Germany.
Engineering
- 20 Oct 2026Embedding models for RAG: how to choose one before you indexVector size, context window and price decide which embedding model to index a RAG corpus with. The open-weight options run in France, and how to pick one.
- 20 Oct 2026Self-hosting an LLM or paying per token: where the break-even really isA rented H100 costs the same whether it serves one token or a billion. The break-even against per-token prices, worked out with real October 2026 prices.
- 20 Oct 2026AI API aggregators: one key for many models, and the question of who runs themAn AI API aggregator puts many models behind one OpenAI-compatible key. How OpenRouter, Together AI, Groq, Fireworks AI and Evanesca differ.
- 6 Oct 2026How to choose an AI API aggregator for your companyFive questions to ask an AI API aggregator before you sign — jurisdiction, data kept, billing, catalogue, compatibility — and how five providers answer them.
- 20 Oct 2026Open-source AI gateways and inference servers you can host yourselfLiteLLM, vLLM, SGLang, Ollama, TGI, Xinference: what each self-hosted tool does, what running it costs, and when a managed EU API is the simpler route.
- 20 Oct 2026OpenRouter alternatives and competitors: what each one changesOpenRouter alternatives and competitors compared on what makes teams leave: jurisdiction, data handling, fees, speed or fine-tuning. Dated and sourced.
Comparisons
- 6 Oct 2026Evanesca vs ChatGPT: what changes for your conversationsChatGPT has OpenAI's own models and more features. Evanesca is an AI agent and chat run in Europe that keeps no copy of your chats. Both sides, sourced.
- 6 Oct 2026Evanesca vs Mistral Vibe, formerly Le Chat: two French AI chats, two ways of handling your wordsMistral Vibe, formerly Le Chat, is French and feature-rich, but keeps your chats and trains on them by default. Evanesca keeps no copy. Compared.
- 6 Oct 2026Evanesca vs Lumo by Proton: two private AI chats, one Swiss and one FrenchLumo, by Proton, keeps no logs and syncs an encrypted history. Evanesca, a French AI agent and chat, runs its models in Europe, paid per use. Compared.
- 6 Oct 2026Evanesca vs OpenRouter: one router for every model, or models run in EuropeOpenRouter offers far more models, often for less. Evanesca runs its models in Europe with European operators. Price, data and law, side by side.
- 13 Oct 2026Evanesca vs OrcaRouter: a zero-markup router from Singapore, or models run in EuropeOrcaRouter is cheaper and far broader. Evanesca runs every model in Europe, each at a named operator. Price, prompts, region and law, side by side.
- 6 Oct 2026Evanesca vs Together AI: the same open models, other prices, other defaultsTogether AI is cheaper on most open models both serve, and fine-tunes. Evanesca runs its models in Europe and stores no prompt at all.
- 6 Oct 2026Evanesca vs Groq: raw speed, or models run in EuropeGroq is built for speed and has a free plan. Evanesca runs its models in Europe and keeps no prompt. GPT-OSS prices, data and law compared.
- 6 Oct 2026Evanesca vs Fireworks AI: fine-tuning and speed tiers, or models run in EuropeFireworks AI keeps no prompts by default and adds fine-tuning and speed tiers. Evanesca runs its models in Europe. Data, region and price compared.
- 6 Oct 2026Evanesca vs Venice AI: encrypted privacy modes, or models run in EuropeVenice AI offers encrypted and enclave privacy modes, often at lower prices. Evanesca runs its models in Europe with European operators. Side by side.