Skip to main content
Building your own agent instead of using an off-the-shelf runtime? AIsa plugs in at two levels: an OpenAI-compatible (and Anthropic-compatible) model endpoint for inference, and one instruction file that teaches your agent the full AIsa capability layer. AIsa endpoint:

Prerequisites

Before you start, make sure you have:
  • An AIsa API key (console.aisa.one/api-keys)
  • An agent codebase — any language or framework that can speak the OpenAI or Anthropic API

1. Point your agent’s LLM calls at AIsa

Keep the key in an environment variable:
The same key reaches every model in the catalog — swap the model string to switch providers, no new SDK or key needed. Agent frameworks (LangChain, Vercel AI SDK, LlamaIndex, …) work the same way: set the OpenAI-compatible base URL to https://api.aisa.one/v1.

2. Teach your agent the AIsa capability layer

AIsa publishes its agent-facing instructions — endpoints, Skills, and usage rules — as a single markdown file:
Two ways to use it:
  • System prompt: fetch the file at startup and include it in your agent’s system prompt or tool context, so the agent knows how to call AIsa’s search, data, and media APIs.
  • On demand: if your agent can browse or fetch URLs, give it the standard setup prompt:
Try it end to end:

Quick reference

Troubleshooting

Agent Quickstart

The agent-facing instruction file this tutorial builds on.

Authentication

API key creation, rotation, and storage best practices.

Model Catalog

Every model ID, context window, and endpoint your agent can call.