Skip to main content

Platform Agnostic Integration

Spec27 connects to many different platforms and services through a single, simple agent contract, so teams can evaluate agents they build themselves, agents that call internal APIs, and agents backed by third-party services — all in the same workflow.

Three ways to connect an agent​

You rarely start from a blank file. Spec27 offers three paths, and they all produce the same kind of agent:

  • Registry integrations — copy a prebuilt integration for a popular platform (OpenAI, Gemini, LangGraph, Botpress, Intercom, Zendesk, and more) and supply your credentials.
  • Agent Builder — describe the agent in plain language and have Spec27 generate the code.
  • Writing the code yourself — author the agent directly when you need full control.

The underlying contract​

Under all three paths, an agent is JavaScript that receives an input, calls the target system, and returns the output to be scored:

return async function process(input) {
// call the target system
// return the final output Spec27 should score
}

Because the contract is small, the same model wraps many different backends — internal services behind an API, third-party model or agent providers, and vendor tools that can be called from JavaScript.

What "platform agnostic" means here​

It does not mean every system looks the same internally. It means the integration surface is flexible enough that different systems can be wrapped and evaluated through the same product workflow — so teams do not need a separate evaluation system for each platform they assess.

Example: connecting a new vendor-backed agent​

Suppose a team wants to evaluate a vendor-backed assistant alongside an internally built one. If the vendor is in the registry, they copy that integration and add their credentials; otherwise they describe it in Agent Builder or write a small JavaScript wrapper. Either way, the new agent connects to the same specifications and eval workflows already used elsewhere in the project.