Summarizer

Building block for the Web’s Built-in Summarizer API. Skeleton extraction, sentence-boundary trim, streaming, and pluggable result caching. The React adapter lives at @web-ai-sdk/summarizer/react; see useSummarizer.

Usage

import { summarize } from "@web-ai-sdk/summarizer";

const result = await summarize({
  language: "en",
  article: document.querySelector("article") ?? undefined,
  createOptions: { type: "key-points", length: "short" },
  onUpdate: (text) => render(text),
});

console.log(result.summary, result.cached);

Pass article to let the wrapper extract a skeleton (title + description + every h1-h4 and <strong> / <b> inside the article); or pass text directly to skip skeleton extraction. The input is trimmed to a sentence boundary (default cap: 3000 chars) so the model never sees half-cut sentences.

How it works

1. Build a skeleton. For long posts the high-signal content drops the input from thousands of chars to a few hundred and produces a tighter summary. Falls back to the trimmed body when the skeleton is too thin (minSkeletonChars, default 200).
2. Trim to a sentence boundary. So the model never sees a half-cut sentence (default cap: 3000 chars).
3. Cache Summarizer.create() sessions by JSON-stringified options. First post pays the ~1-3s cold start; later same-config calls reuse the warm session.
4. Stream summarizeStreaming() when the instance supports it, falling back to one-shot summarize(). Cleaned chunks are pushed to onUpdate as they arrive.
5. Optionally cache the final summary when you pass a cache (e.g. createSessionStorageCache()). Off by default; opt in for revisits in the same tab to render instantly.

Two caches, two purposes

Session cache (internal, in-memory, always on): a Map<stringifiedOptions, Summarizer> so the second call with the same { language, sharedContext, createOptions } reuses the warm Summarizer.create() session. Cold-start is ~1-3s; warm sessions are sub-second. Cleared on full page reload.

Result cache (opt-in): off by default; every call hits the model. Pass a cache (anything matching { get, set }) to memoize the final summary string by cacheKey (or a default key derived from the inputs).

import { summarize, createSessionStorageCache } from "@web-ai-sdk/summarizer";

// Off by default; every call hits the model.
summarize({ language: "en", article });

// Per-tab caching via sessionStorage
summarize({
  language: "en",
  article,
  cache: createSessionStorageCache(),
});

// Persistent caching across tabs
summarize({
  language: "en",
  article,
  cache: createSessionStorageCache({ storage: localStorage, prefix: "post-summary:" }),
});

result.cached tells you which path served the response, so you can render a “From cache” hint or skip a re-fetch.

Streaming

When the underlying Summarizer instance exposes summarizeStreaming(), the library uses it and pushes cleaned chunks to onUpdate as they arrive (cumulative buffer, not deltas). Otherwise it falls back to one-shot summarize(). Either way, result.summary is the final cleaned text.

Output normalization

cleanSummary strips wrapping quotes / whitespace and collapses internal whitespace; applied to every result regardless of type. Anything beyond that — e.g. trimming the trailing period from a type: "headline" result so it reads as a label rather than a sentence — is your concern. A one-line regex after the call covers the headline case.

Language support beyond en/es/ja

The Web’s Built-in Summarizer (Chrome 138+, Edge 138+) only accepts expectedInputLanguages / outputLanguage for ["en", "es", "ja"]. Pass any other language as language: "pt" and the library omits those hints. The model still runs, but you steer it via sharedContext instead:

summarize({
  language: "pt",
  article,
  sharedContext: { pt: "Resuma o artigo em português em 2-3 frases curtas." },
});

When Chrome adds more accepted languages, pass them explicitly via supportedLanguages: ["en", "es", "ja", "pt"] and the hints fire through.

Errors and unavailability

The vanilla summarize() throws SummarizerUnavailableError when the API is missing. Callers branch explicitly:

import { summarize, SummarizerUnavailableError } from "@web-ai-sdk/summarizer";

try {
  const result = await summarize({ language: "en", article });
} catch (err) {
  if (err instanceof SummarizerUnavailableError) {
    return;
  }
  throw err;
}

AbortSignal is supported. Aborting mid-stream resolves cleanly; an opt-in result cache is not written for aborted runs.