Build a Telegram video downloader bot with grammY

grammY only; fetch is built into Node 18+. There is no scraping stack to install and no browser to keep alive.

npm install grammy

One GET with the key in the Authorization header. AbortSignal.timeout keeps a slow source from holding your handler. The endpoint detects the platform from the URL, so the bot does not branch on the domain.

const API = "https://api.saveapi.org/v1/download";

async function resolve(link) {
  const res = await fetch(API + "?url=" + encodeURIComponent(link), {
    headers: { Authorization: "Bearer sk_live_xxxx" },
    signal: AbortSignal.timeout(30_000),
  });
  return res.json();
}

message::url matches any message carrying a URL entity; ctx.entities("url") gives you the text. replyWithVideo and replyWithPhoto take the media URL directly — Telegram fetches it. Check medias[0].type: a slideshow or a carousel is images, not a video.

import { Bot } from "grammy";

const bot = new Bot(process.env.BOT_TOKEN);

bot.on("message::url", async (ctx) => {
  const link = ctx.entities("url")[0].text;
  const data = await resolve(link);

  if (!data.success) {
    return ctx.reply(EXPLAIN[data.error.code] ?? "Try another link.");
  }

  const media = data.medias[0];
  return media.type === "video"
    ? ctx.replyWithVideo(media.url, { caption: data.meta?.title ?? "" })
    : ctx.replyWithPhoto(media.url);
});

bot.start();

Users paste private posts and dead links all day. Map the stable codes to sentences: a bot that says "That account is private" gets a different link, a bot that says "error" gets an angry message. Private and missing cost 0.1 credits, not the full 1.5.

const EXPLAIN = {
  PRIVATE_CONTENT: "That account is private — I can only fetch public posts.",
  MEDIA_NOT_FOUND: "That post has no media, or it was deleted.",
  UNSUPPORTED_PLATFORM: "I do not know that site yet.",
  QUOTA_EXCEEDED: "Out of credits.",
  UPSTREAM_TIMEOUT: "The source was slow. Send the link again.",
};

After the first send, Telegram gives you a file_id; re-sending it uploads nothing and costs no credits. The media URL is the wrong thing to store — CDN links are signed and expire within hours.