Start free on your own infrastructure. Add the maintained terms feed and audit evidence when you go to production, and enterprise identity and support when you scale.
Free
For evaluating Wayreeve on your own infrastructure.
Talk to us
For a platform team running regulated workloads in production.
Talk to us
For organisations with several teams and strict audit duties.
One endpoint for OpenAI, Anthropic, Bedrock, Azure OpenAI, Vertex AI and your own GPUs. Load balancing, failover and retries that never leave your permitted routes. Deadline-aware batch planning.
An MCP gateway per project: an approved tool catalog, per-tool permissions and rate limits, approvals for sensitive actions, poisoning and rug-pull checks. Your REST APIs become tools.
Agents are named identities that act on behalf of signed-in users, never beyond what that user may do. Agent-to-agent tasks pass the same checkpoint.
Checks on prompts, responses and tool results, including prompt injection. Bring your own: Presidio, Bedrock Guardrails, Azure Content Safety, Google Model Armor, OpenAI Moderation.
Token limits and budgets by team, project and key, including traffic on your own provider keys. Alerts and hard caps. A response cache where your data rules allow it.
Single sign-on with OIDC and SAML, SCIM provisioning from Okta and Entra, roles, and short-lived credentials for workloads.
Tamper-evident decision records, streamed to Splunk, Datadog, S3 or syslog. Runs multi-replica on Postgres, installed with Helm, including fully disconnected sites.
No. Your providers bill you directly. Wayreeve is licensed separately.
Requests and prompts stay in your own environment. Wayreeve Cloud handles licences, fleet health and the terms feed.
Inference keeps running. Licensed features turn off after a grace period.
With a pilot on your own infrastructure. We agree success criteria first, then talk about the right edition.