For developers
Everything Raytio does is an API, and most of it is also an MCP tool - so it composes into your product whether the caller is your backend, your frontend, or your agent. This page covers both: using Raytio as building blocks, and building agentic systems on the worker platform.
Request identity, biometric, screening and registry checks; receive signed, machine-readable results you can verify independently against our published public key.
Request exactly the fields you need; decrypt submissions with your organisation's keys; nothing readable server-side. Webhooks fire on submission, and an SDK decrypts to JSON, CSV or PDF.
Parties, catalogues, price lists, quotes, orders, subscriptions, usage events, projects - REST endpoints with OpenAPI docs, backed by row-level tenant isolation.
Connect an agent to the platform directly; the same authorisation model applies to it as to your staff, down to field level.
Multi-tenant with database-level isolation, multi-organisation, i18n and multi-currency, entitlements, workflow, batch jobs and a schema-driven admin UI - the enterprise machinery is already built.
Fully managed, or self-hosted in your own AWS account - provisioned entirely from infrastructure as code, so a customer deployment is a repeat of a process we run continuously ourselves.
A complete development platform
Three properties remove most of the boilerplate from a typical build:
The query shape, the visual layout - lists, forms, Kanban boards - the menus and the permissions are all driven from the schema and the database. Extend the data model and the whole interface follows, without custom screens.
Batch jobs can call any HTTP endpoint - your code, on your infrastructure, scheduled, retried and monitored by the platform's job system.
Typed extensible attributes and attachments on every record - add arbitrary data, attributes and objects, and they surface through the UI in a cohesive manner.
Operate - the platform in depth → · Foundation platform services →
Building agents?
The worker platform gives your agents what production systems demand: identity, credentials, guardrails, evaluations, and cost attribution.
Define role, system prompt, model preferences and tool allowlists once; resolve a worker's entire runtime config in a single call.
Agents request credentials from named humans and receive them end-to-end encrypted, time-boxed and revocable - never pasted into prompts or config.
Every dispatch traced with steps, artifacts, tokens and dollar cost - aggregated by worker, role, provider and model.
Worker management → · Agent credentialling → · Agents pillar overview →
Docs at docs.rayt.io. API reference at api-docs.rayt.io. Start with the free tier.
Deployment & hosting → · Authorization & access control → · Foundation platform services →
Verified customers, protected data, and agents you can hold to account.