Who is it for?¶

Ikanos serves several different capabilities, and each one matters to a different kind of reader. The same engine looks like an AI integration layer to one team and an API governance layer to another. This page helps you find the framing that fits you, then points you straight at the use cases and tutorials that build on it.

The one-sentence version: Ikanos reads one YAML file describing a slice of your business, connects it to the HTTP/REST APIs you already have, and serves that same slice to AI agents, web apps, and partners over several protocols at once — no code to write or compile.

You're building AI agents¶

You want your AI assistant or agent to reach real business systems — safely, without hand-writing and maintaining a pile of glue code.

What Ikanos can do for you:

• Give agents governed tools without writing handlers. Describe an operation in YAML and Ikanos serves it as an MCP tool — the open standard AI clients speak — with no Python or Java handler to build and trust.
• Send the model exactly the context it needs. Trim and reshape large API responses down to the few fields a task actually uses, so prompts stay relevant and cheap.
• Compose context from several systems at once. Join a customer, their orders, and their open tickets from three different APIs into one clean result the agent receives in a single call.
• Stay in control of what the AI produced. The capability is plain, reviewable YAML — a readable diff in a pull request that can be linted with Polychro before it ever runs.

You own a sprawl of APIs, microservices, or a monolith¶

You're a platform or integration team, and the systems you expose are messy, inconsistent, or hard for other teams to consume.

What Ikanos can do for you:

• Elevate an existing API into something easier to consume. Put a stable, business-oriented contract in front of vendor-native or legacy endpoints — and even fix their semantics (turn a read-only POST query into a cacheable GET).
• Rightsize microservices or a monolith. Hide service fragmentation behind one curated namespace, or carve focused capabilities out of a broad monolith so each consumer gets only what it needs.
• Convert formats for free. Upstream returns XML, CSV, Avro, Protobuf, or Markdown? Ikanos normalizes it to clean JSON on the way out.
• Govern it in CI. Conventions — naming, security, path rules — are enforced by Polychro linting before anything ships.

You want one source of truth for every kind of client¶

You're a product or partner-facing developer, and the same domain has to be reachable by web apps, partner integrations, and AI agents — without maintaining three separate implementations.

What Ikanos can do for you:

• Write the domain once, expose it many ways. A single capability is served as REST for conventional clients, MCP for AI agents, and Skill for agent skill catalogs — from the same YAML.
• Keep contracts in sync automatically. Define a reusable domain function once (an aggregate) and project it onto each protocol; no drift between your REST API and your MCP tools.
• Interoperate with what you already have. Import an existing OpenAPI document to bootstrap a capability, and export your REST surface back out as OpenAPI for downstream consumers.

Not sure which one you are?¶

Most teams are more than one of these at the same time — that's the point. Pick the framing closest to your immediate goal, run its tutorial, then come back and explore the others. The Use Cases catalog lists every concrete scenario, and the Features page is the exhaustive reference.