Skip to content

Infrastructure you can build on.

Dianthan operates the backend that products depend on (identity, payments, data, messaging, model access and observability) as one managed platform, and runs an engineering practice that takes products to production on it.

The layer beneath the product.

Much of what software needs before it can serve its first customer is the same every time. We operate that layer, and we help teams build on it.

Platform

Managed systems for identity, billing, usage metering, data, messaging, automation, observability, release control and model access. Each has its own API; together they share one identity and one audit trail.

The platform

Infrastructure

Requests are served from a global edge network. Data is replicated, backed up continuously, and isolated per product by design. Availability and incidents are reported publicly.

How it runs

Engineering

Engineers who take products to production, build features and model integrations, and look after them afterwards, each under a written scope and a published starting price.

Services

Every model, one endpoint.

Call any major model through one API your libraries already speak. Fallbacks keep a request alive when a provider isn't, and every call is priced and attributed to the customer who made it, with a cap so no single account can run up your bill. Because the same platform takes their payments, you see what each customer pays against what they cost.

Billed at the provider's price plus 8%.

TypeScript
const ai = new OpenAI({
  baseURL: "https://api.dianthan.com/v1/ai",
  apiKey: process.env.DIANTHAN_KEY,
});

await ai.chat.completions.create({
  model: "anthropic/claude-sonnet-5.5",
  models: ["openai/gpt-5"],   // fallback
  messages,
  user: session.userId,         // cap and cost per customer
});

Engineers, alongside the platform.

Software is now written faster than it is made dependable. Our engineers close that gap: they take products to production, build what is missing, and stay on to look after it. Each engagement has a written scope and a published starting price, and using the platform is never a condition.

  • Production readiness

    For a product that works but is not yet ready for paying customers. We review it, close the gaps that matter, and hand back a written report.

    From $3,000
  • Product engineering

    Web applications, APIs and internal tools, designed and built with your team, in milestones you can use as they land.

    Milestones from $5,000
  • AI integration

    Bringing models into a product so that they are dependable and their cost is understood: retrieval, tool use, evaluation, and a cost on every call.

    Milestones from $5,000
  • Operations and care

    Monitoring, upkeep and incident response for systems we have built or reviewed, with a fixed allowance for small changes each month.

    From $400 a month

How it is built.

A few decisions, made early, that everything else rests on.

  • Isolation by construction

    Every record carries the product it belongs to, and every query is scoped by it. Cross-tenant access is prevented at the data layer and checked by tests on every release.

  • One control plane

    Products, environments and teams sit under one account. Each is isolated; all of them are visible and governed in one place.

  • Everything attributed

    Every change, whether made by a person, a server or an agent, is written to an append-only audit log with the credential that made it.

  • Open interfaces

    HTTPS and JSON throughout, standard OAuth sign-in, signed webhooks. No proprietary client is required, and your data can be exported in documented formats at any time.

Operable by people and by agents.

In preview

Everything in the dashboard is also available through the API, the command line and an MCP server, with scoped, expiring credentials. Changes made by an agent are attributed and auditable like any other.

Build on it, or have us build it.

Free while you build, one plan when you launch. Or tell us what you are building.