Sovereign AI on every surface
Web, desktop, mobile, and command line — each a focused way to use the AgentykLM lineup. Every app switches seamlessly between edge and cloud inference, and all of it stays in EU jurisdiction.
AgentykCloud
Web app
Privacy-first AI chat in your browser. The server is stateless — conversations live in your browser, never in a database of ours. Broad model selection, audio transcription, optional web search.
- Stateless server, no conversation database
- 100% Dutch hosted
AgentykChat
Desktop client
Native client for Windows, Mac, and Linux. Privacy-sensitive tasks run fully offline on local inference, with seamless cloud switching when a task needs a bigger model.
- Full offline capability
- Local inference + cloud fallback
AgentykMobile
iOS & Android
On-device AI for field workers, technicians, and sales teams. Runs the smaller AgentykLM tiers without an internet connection — edge-first by design.
- On-device models, no internet required
- Field-ready UX
AgentykCode
CLI tool
Command-line interface for code generation, review, and debugging. Integrates with existing IDEs and runs against local or cloud models on the same key.
- Code generation & review
- IDE integration
Beyond the apps
The same lineup, available to build on directly.
AgentykAPI
One OpenAI-compatible API for chat, transcription, embeddings, and reranking — a single agk_ key across all ten tiers.
Agentyk Knowledge
Enterprise grounding on a verifiable atomic knowledge graph — every answer traces to the exact source sentence.