ChatAPI

The primary API for text-based AI interactions.

Methods

get(request, ctx?)

Non-streaming chat completion:

const response: Response = await ai.chat.get(
  { messages: [{ role: 'user', content: 'Hello' }] },
  ctx?
);

stream(request, ctx?)

Streaming chat completion:

for await (const chunk: Chunk of ai.chat.stream({ messages }, ctx?)) {
  if (chunk.content) process.stdout.write(chunk.content);
}

createExecutor()

Create a standalone executor function:

const execute: Executor = ai.chat.createExecutor();
const response = await execute(request, ctx, model);

createStreamer()

Create a standalone streamer function:

const stream: Streamer = ai.chat.createStreamer();
for await (const chunk of stream(request, ctx, model)) { /* ... */ }

Capability Detection

ChatAPI automatically detects required capabilities from the request:

Condition	Capability
Always	chat
Images in messages	vision
Audio in messages	hearing
request.reason set	reasoning
responseFormat: 'json'	json
responseFormat: { type: zodSchema }	structured
request.tools present	tools
Stream mode	streaming

Request Type

See Request in @aeye/core types.

Response Types

• get() returns Response
• stream() yields Chunk