Back to vendors
n

nexos.ai

Also known as: Nexos, Nexos AI

Visit site
Agent infrastructureindependentVerified 2026-07-08

Enterprise AI control layer from the Nord Security founders pairing an AI Gateway, one secure endpoint routing more than two hundred models with caching, fallbacks, RAG, and spend controls, with a governed AI Workspace where employees chat and build reusable agents under guardrails and full logging.

nexos.ai was founded in late 2024 in Vilnius, Lithuania by Tomas Okmanas and Eimantas Sabaliauskas, the co founders who bootstrapped Nord Security, the three billion dollar cybersecurity company behind NordVPN, over twelve years. They emerged from stealth with eight million dollars led by Index Ventures with Creandum, Dig Ventures, and angels including Datadog's Olivier Pomel, Klarna's Sebastian Siemiatkowski through Flat Capital, Supercell's Ilkka Paananen, and Wix's Avishai Abrahami, then closed a thirty million euro Series A in October 2025, roughly thirty five million dollars, co led by former Sequoia partner Matt Miller's Evantic Capital and Index, just months after launch. The thesis, in Okmanas's words: employees uploading sensitive data to LLMs is the biggest corporate data leak in the making, and the answer is a Switzerland for LLMs sitting neutrally between teams and AI systems rather than a ban.

The platform has two halves. The AI Gateway gives developers plug and play API access to more than two hundred models from providers including Anthropic, OpenAI, Google, Mistral, and Meta through a single endpoint, with smart routing, load balancing, intelligent caching to cut cost and latency, smart fallbacks during model outages, RAG to ground responses in company documents, spend controls, bring your own keys, and private model hosting for sensitive workloads. The AI Workspace gives employees governed chat and no code reusable agents with customizable guardrails, role based access control, side by side model comparison, any file format, built in web search, and detailed logging with trace visibility, with agents connecting to Google Workspace, SharePoint, Slack, GitHub, and Zendesk. Full interaction logs export into security and identity systems, the infrastructure is SOC 2 and ISO 27001 certified and GDPR compliant, hosting is in Europe with regional data residency, and zero data retention can be enabled at the LLM level. Disclosed customers include Nord Security, NordVPN, Lithuanian retailer Pigu.lt, fintech unicorn Payhawk, and Tesonet portfolio companies.

nexos.ai fits European and regulated enterprises consolidating shadow AI into one governed layer, and competes across the gateway field with Portkey, LiteLLM, and OpenRouter while the workspace side meets Dust and Glean. It is a weaker fit for buyers needing published rates or self hosted deployment of the platform itself.

Vendor details

Canonical URL

https://nexos.ai

Category

Agent infrastructure

Subcategory

Enterprise AI gateway and governed AI workspace

Funding status

Independent, founded in late 2024 in Vilnius, Lithuania by Tomas Okmanas and Eimantas Sabaliauskas, the Nord Security co founders. Emerged from stealth with eight million dollars led by Index Ventures with Creandum, Dig Ventures, and angels including Olivier Pomel, Sebastian Siemiatkowski through Flat Capital, Ilkka Paananen, and Avishai Abrahami, then closed a thirty million euro Series A, about thirty five million dollars, in October 2025 co led by Evantic Capital and Index Ventures. Disclosed customers include Nord Security, NordVPN, Pigu.lt, Payhawk, and Tesonet portfolio companies.

Company status

independent

Use cases & customers

Primary use cases

Consolidating shadow AI into one governed access layerModel routing, caching, and failover through a single endpointGoverned employee chat and no code reusable agentsAI spend visibility and cost control for finance

Target customers

European and regulated enterprisesSecurity and compliance leaders governing AI adoptionPlatform teams consolidating model accessMid market companies replacing scattered AI tools

Deployment options

Cloud

Integrations

The Gateway exposes more than two hundred models from Anthropic, OpenAI, Google, Mistral, Meta, and others through one API endpoint with bring your own keys. Workspace agents connect without code to Google Workspace, SharePoint, Slack, GitHub, Zendesk, and other work apps, with built in web search, and full interaction logs export into enterprise security and identity management systems.

In practice

Employees across a company paste sensitive data into a dozen public AI tools. nexos.ai replaces the sprawl with one workspace where every interaction runs under guardrails, RBAC, and full logging exported to the security stack.

A development team hard wires one model provider and eats an outage. The Gateway routes through a single endpoint with smart fallbacks and load balancing, so traffic shifts automatically and caching trims the bill.

An online retailer needs four million product descriptions optimized. A reusable agent grounded in company data does per item work the company reports as ninety nine point eight percent cheaper than manual effort.

Capability coverage

9.5 / 14 capabilities · 68%

Integrations & Tool CallingMore than two hundred models accessible through one endpoint with bring your own keys, and Workspace agents connect without code to Google Workspace, SharePoint, Slack, GitHub, Zendesk, and other work apps with built in web search, nexos.ai homepage FAQ retrieved 2026-07-08 Full
Workflow OrchestrationThe Gateway orchestrates model traffic with routing, load balancing, and fallbacks, and the Workspace offers reusable agents, though multi step workflow orchestration is not documented, funding announcement and coverage retrieved 2026-07-08 Partial
Knowledge Grounding & RAGThe AI Gateway provides Retrieval Augmented Generation capabilities to ground responses in company documents, documented as a core Gateway feature, nexos.ai funding announcement and EU-Startups coverage retrieved 2026-07-08 Full
Human Oversight & GuardrailsCustomizable guardrails and role based access controls govern Workspace use with policy applied at the platform edge, though human approval workflows are not documented, EU-Startups coverage and Forbes profile retrieved 2026-07-08 Partial
Security, Identity & GovernanceSOC 2 and ISO 27001 certified infrastructure, GDPR compliance, SSO and RBAC, zero data retention at the LLM level, a public Trust Center, and data never used for training without permission, nexos.ai homepage FAQ retrieved 2026-07-08 Full
Observability & AuditabilityFull interaction logs export into company security and identity management systems, with detailed logging and trace visibility in the Workspace and spend tracking finance can follow, Forbes profile and EU-Startups coverage retrieved 2026-07-08 Full
Memory & State PersistenceChat history persists and travels with users across devices, though an agent memory architecture is not documented, nexos.ai homepage retrieved 2026-07-08 Partial
Deployment & Data ResidencyThe platform and most models are hosted in Europe with regional data residency and private model hosting within nexos.ai infrastructure for sensitive workloads, though customer self hosting of the platform is not documented, nexos.ai homepage FAQ retrieved 2026-07-08 Partial
Prebuilt Agents, Templates & PacksThe Workspace ships governed, ready to use experiences and customers build reusable agents in minutes, though a prebuilt agent or template catalog is not documented, Forbes profile and nexos.ai homepage retrieved 2026-07-08 Partial
Triggers & Channel CoverageAccess spans the web Workspace, a mobile experience, and the developer API, but an event driven trigger framework is not documented, nexos.ai homepage retrieved 2026-07-08 Partial
Model Flexibility & RoutingThe platform is explicitly model agnostic across more than two hundred models from Anthropic, OpenAI, Google, Mistral, and Meta with smart routing, load balancing, fallbacks, and free per task model choice, nexos.ai homepage FAQ and funding announcement retrieved 2026-07-08 Full
APIs, SDKs & MCP ExtensibilityPlug and play API access orchestrates all models through one endpoint, bring your own keys is supported, and no code agent building extends the Workspace, though no MCP surface is documented, funding announcement and homepage FAQ retrieved 2026-07-08 Full
Testing, Debugging & OptimizationTeams can compare models side by side in the Workspace, though a testing or evaluation suite is not documented, EU-Startups coverage retrieved 2026-07-08 Partial
Browser & Computer UseNo browser or computer use capability is described on retrieved pages, nexos.ai retrieved 2026-07-08 Unable to verify

Recent platform changes

No recent material changes tracked yet.

Pricing

Monthly and annual plans for individuals and teams with no published rates; enterprise pricing is customized through sales

monthly or annual subscription plans for individuals and teams plus customized enterprise contracts, per the company FAQ; rates not published

Contact onlyMedium variable cost

What is public

The tier structure is public per the company FAQ: flexible monthly and annual plans for individuals and teams, with annual plans positioned as best value, and customized enterprise pricing. No dollar rates are retrievable from fetched pages.

Billing mechanics

Plan based subscriptions on monthly or annual billing per the FAQ. Bring your own keys is supported, letting customers keep existing AI provider plans, in which case model consumption bills through the provider while nexos.ai supplies the governance layer. Enterprise agreements are customized through sales.

Cost watchouts

Gateway routed model consumption scales with workload, and bring your own keys moves that spend to provider bills that sit outside the nexos.ai subscription.

Variable cost rationale

No public rates exist. The FAQ describes plan based subscriptions rather than pure usage billing, and the Gateway includes cost control features like intelligent caching and spend controls that push against usage growth, but underlying model consumption routed through the Gateway remains workload driven, especially with bring your own keys where provider costs bill separately.

Additional watchouts

With bring your own keys, total cost splits between the nexos.ai subscription and the underlying model provider bills, so budget both. Private model hosting for sensitive workloads is presumably priced separately, worth confirming in enterprise scoping.

Sales call required

Mixed (some tiers require a call)

Free / trial

No free tier or trial is documented on retrieved pages

Key ambiguities

Whether team plans price per seat, by usage through the Gateway, or both, and how private model hosting is charged.

Missing data

No published rates for individual, team, or enterprise plans, no seat or usage quotas, and no minimums were retrievable.

Verified 2026-07-08

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.