Chatbots answer; gateways forward API calls. Nexevo is a vertical AI platform: your workspace carries private memory, project knowledge, and custom-distilled skills, and an autonomous agent does multi-step work inside that context. The model layer — routing, fallbacks, upgrades — is our job, not another system for your team to operate.

Simple, model-agnostic rates — 100 credits equals one US dollar. Chat is priced per million tokens by context length, images per image, video per second, speech per thousand characters. Which model served the request never changes your price — picking the right one is our problem. Memory capacity tiers and enterprise terms come with your design-partner agreement.

A credit is Nexevo's unit of usage — 100 credits equals one US dollar. Each modality is a flat per-unit rate: chat per million tokens, images per image, video per second, speech per thousand characters. The model that serves a request never changes the price — picking the right one is our problem.

A routed pool of 15+ frontier models across the leading providers, plus self-hosted inference for memory embedding and reranking. You never pick a model — the router selects the right one per task, and which one served it never changes your price.

Yes. API-key services — embedding, reranking, transcription and more — can run on keys you bring, connected per workspace and self-served in Settings. Your key never subsidizes another tenant's usage.

Your prompts, outputs, and knowledge files are never used to train foundation models, and upstream providers receive requests under no-training terms. Vertical skills are built from anonymized, aggregated patterns — never from your identifiable content, and your private data never appears in anyone else's workspace.

Primary infrastructure runs in Singapore with strict per-workspace isolation across storage, memory, and files. Export or delete your workspace data anytime.

Credits from a one-off pack are valid for one year from purchase. Credits that come with your subscription — monthly credits, the daily refill, and the signup bonus — refresh each cycle while your plan is active.

Unused prepaid balances are refundable for 12 months, and first-time subscriptions carry a 14-day full-refund cooling-off period. Usage already metered — chat, images, video, speech — is non-refundable, since upstream providers bill us for it. The full policy (processing times, service-outage credits, chargebacks) is on the refund policy page.

Talk to sales. We onboard a small number of design partners per vertical: a scoping call, a co-build phase with your experts, then a team-wide rollout — typically weeks, not quarters.