Use one governance platform when models and agents need different access, privacy, spending or deployment rules. Start with the situation closest to yours, then see the control that resolves it.
Free tier on the Hosted Platform. No card required.
Give each team a list with its approved models, tools, data access, privacy rules and budget, while every application keeps one ARC integration.
Run one deployment with a separate model, tool, data and policy boundary for each customer, with no shared pool.
Give each agent a list that bounds model, tool, approved data access and spend outside its prompt. ARC does not control the agent's reasoning or runtime.
Keep the underlying material inside your boundary while an approved external model receives only the shielded context its step requires.
Self-hosted Licence only
Run ARC, models, tools and connected storage in your VPC, on your hardware or fully air-gapped. Request data stays inside your network boundary.
Apply budgets by organisation, list and person. ARC can prefer lower-cost approved routes or block work at a hard cap without weakening capability, privacy or access.
Set the next approved path in advance. Applications keep the same ARC interface while a local rule-based path can continue routing through an outage.
FenxLabs can assess, deploy and operate the agreed ARC and model environment in your infrastructure or ours, while your organisation keeps the decisions and accountability.
A solo developer and a platform team integrate the same way: one OpenAI-compatible base URL and one key. The approved models, tools and policy behind that key are configuration.
Start free with one OpenAI-compatible integration. Talk to an engineer if access policy, data boundaries or deployment need to be settled first.