Live explainer

Ask any question. The model's answer streams back token by token — the same typewriter UX that makes chat apps feel responsive even when the total latency is similar to a one-shot call.

Run

bash bootstrap-secrets.sh              # reads ../../../../.env, writes secrets/
docker compose up --build              # default: PROVIDER=openai

Open http://localhost:3000.

To run against Gemini instead:

PROVIDER=gemini docker compose up --build

What's where

• backend/ai_openai.py — run_stream yields tokens, run accumulates
• backend/ai_gemini.py — same shape, other SDK
• backend/main.py — FastAPI: /api/ai (final) + /api/ai/stream (SSE)
• frontend/app/page.tsx — reads the SSE stream and types output as it arrives

Stop

docker compose down

unzip live-explainer.zip
cd live-explainer
bash bootstrap-secrets.sh   # one-time: pulls API keys into ./secrets
docker compose up --build   # default provider: openai
# or:  PROVIDER=gemini docker compose up --build

Los mismos módulos que ejecuta el botón Run. El proyecto completo (frontend, Dockerfile, compose) está en el ZIP, pestaña README.

backend/ai_openai.py

"""Week 3 - Showcase 1 (OpenAI): live answer to a question, streamed.

The streaming pattern itself is the lesson — `run_stream` yields tokens
as they arrive from the server, which the FastAPI app forwards over SSE
to the browser so the reader watches the answer take shape. `run` is a
thin accumulator over `run_stream`, kept so the website's generic Try-it
widget can call a `def run(input: str) -> str` and get the final text.
"""
from typing import Iterator

from openai import OpenAI

_client = OpenAI()  # reads OPENAI_API_KEY from the environment.


def _prompt(question: str) -> str:
    return (
        "Answer the following question in two to four short paragraphs. "
        "Use plain prose only; no headings, no bullet lists, no code blocks."
        f"\n\n---\n{question}\n---"
    )


def run_stream(question: str) -> Iterator[str]:
    # Cap output length so a runaway answer can't burn the bill.
    # 2048 tokens (~6k chars) is well over the prompt's 2-4 paragraphs.
    stream = _client.responses.create(
        model="gpt-5.4-nano", input=_prompt(question), stream=True,
        max_output_tokens=2048,
    )
    for event in stream:
        if event.type == "response.output_text.delta":
            yield event.delta


def run(question: str) -> str:
    return "".join(run_stream(question))

backend/ai_gemini.py

"""Week 3 - Showcase 1 (Gemini): live answer to a question, streamed.

Same lesson as ai_openai.py — `run_stream` yields tokens as they arrive
from the Gemini server, `run` accumulates for the non-streaming contract.
"""
import os
from typing import Iterator

from google import genai
from google.genai import types

_client = genai.Client(api_key=os.environ["GEMINI_API_KEY"])


def _prompt(question: str) -> str:
    return (
        "Answer the following question in two to four short paragraphs. "
        "Use plain prose only; no headings, no bullet lists, no code blocks."
        f"\n\n---\n{question}\n---"
    )


def run_stream(question: str) -> Iterator[str]:
    # Cap output length so a runaway answer can't burn the bill.
    # 2048 tokens (~6k chars) is well over the prompt's 2-4 paragraphs.
    stream = _client.models.generate_content_stream(
        model="gemini-3.1-flash-lite", contents=_prompt(question),
        config=types.GenerateContentConfig(max_output_tokens=2048),
    )
    for chunk in stream:
        if chunk.text:
            yield chunk.text


def run(question: str) -> str:
    return "".join(run_stream(question))

Archivos del proyecto

• .gitignore
• README.md
• backend/Dockerfile
• backend/ai_gemini.py
• backend/ai_openai.py
• backend/main.py
• backend/requirements.txt
• bootstrap-secrets.sh
• docker-compose.yml
• frontend/Dockerfile
• frontend/app/layout.tsx
• frontend/app/page.tsx
• frontend/next.config.ts
• frontend/package.json
• frontend/tsconfig.json