Stream text

Print a model response incrementally with streamText.

The callback receives normalized StreamEvent values. This example writes only seTextDelta events, so the answer appears on screen while the model is still generating it.

import std/os
import nimgent
import nimgent/providers/openai

let model: LanguageModel = openAI(getEnv("OPENAI_API_KEY")).model("gpt-4o-mini")

echo "calling the model..."
let _: ProviderResponse = streamText(
  model,
  system = "You are a helpful assistant. Answer without using markdown.",
  prompt = "Explain why the sky is blue.",
  onEvent = proc (ev: StreamEvent): bool =
    if ev.kind == seTextDelta:
      stdout.write ev.text
      flushFile(stdout)
    true)

echo ""

Run it from the repository root:

OPENAI_API_KEY=... nim c -r examples/stream_text.nim

View the source example