Start free. Scale as your agents do. No surprise bills when an agent goes rogue.
Perfect for prototypes, side projects, and evaluating Agent Mounts before committing.
For teams running agents in production that need full audit trails and reliable access control.
For organizations that need unlimited mounts, SSO, dedicated infrastructure, and security review.
A mount is a single configured connection between an agent and a data source. One PostgreSQL database = one mount. One S3 bucket = one mount. Different permission scopes on the same database = separate mounts.
Yes — encrypted at rest using envelope encryption, isolated from agent access. Credentials are only used by the mount proxy to fulfill agent requests. They are never returned to the agent or logged.
On Free, requests beyond the limit are rate-limited and return an error. On Pro, we'll alert you and allow a 20% overage grace period. We won't silently charge you.
Yes. Agent Mounts exposes a standard REST API. Works with LangChain, LlamaIndex, AutoGen, custom agents, or anything that can make an HTTP request.
Free plan. No credit card required. First mount in 5 minutes.
Get early access