Ask Content

useAskContent lets users ask natural language questions and receive AI-generated answers grounded in your project’s actual content. The answer streams token by token via SSE, and once streaming completes, a list of source records is provided.

Usage Example

import { useAskContent } from "@sublay/react-js";

function AskBox() {
  const { answer, sources, streaming, loading, error, ask, reset } = useAskContent();

  const handleAsk = () => {
    ask({
      query: "What are the most common questions about billing?",
      sourceTypes: ["entity", "comment"],
      limit: 5,
    });
  };

  return (
    <div>
      <button onClick={handleAsk} disabled={loading || streaming}>Ask</button>
      <button onClick={reset}>Clear</button>

      {loading && <p>Thinking...</p>}
      {answer && <p>{answer}{streaming && "▋"}</p>}

      {sources.length > 0 && (
        <ul>
          {sources.map((s) => (
            <li key={s.record.id}>
              [{s.sourceType}] {s.record.id}
            </li>
          ))}
        </ul>
      )}
      {error && <p style={{ color: "red" }}>{error}</p>}
    </div>
  );
}

Parameters (`ask` function)

string

required

The natural language question to answer.

("entity" | "comment" | "message")[]

Limit the content types searched for context. Defaults to all types.

string

Scope the context lookup to a specific space.

boolean

default:"false"

Only applies with spaceId. When true, the context lookup also covers every space nested under it — the whole subtree at any depth — so you can ask across a community and all of its channels at once. Ignored without spaceId.

string

Scope the context lookup to a specific conversation.

number

Maximum number of source records to retrieve and use as context.

Returns

string

The AI-generated answer. Grows incrementally as tokens arrive during streaming.

ContentSearchResult[]

The source records that were used as context for the answer. Populated after streaming completes.

Show ContentSearchResult properties

"entity" | "comment" | "message"

The type of the source record.

number

Cosine similarity score (0–1) for this source.

Entity | Comment | ChatMessage

The full source record.

boolean

true while the SSE stream is open and tokens are arriving.

boolean

true from when ask() is called until the first token arrives (or an error occurs).

string | null

Error message if the request failed.

(props) => void

Initiates the AI question. Cancels any in-flight stream before starting a new one.

() => void

Aborts any in-flight stream and clears all state.

useAskContent uses the Fetch API with ReadableStream and TextDecoder for SSE. In React Native, install react-native-fetch-api, web-streams-polyfill@^3, react-native-polyfill-globals, and text-encoding, then call the targeted sub-polyfills once at app startup:

import { polyfill as polyfillEncoding } from "react-native-polyfill-globals/src/encoding";
import { polyfill as polyfillReadableStream } from "react-native-polyfill-globals/src/readable-stream";
import { polyfill as polyfillFetch } from "react-native-polyfill-globals/src/fetch";

polyfillEncoding();
polyfillReadableStream();
polyfillFetch(); // last, so it wins over the built-in fetch

Prefer these over the all-in-one polyfillGlobals(), which also pulls in the native react-native-get-random-values and forces a development-client rebuild. Pin web-streams-polyfill to v3 — v4 removed the ponyfill/es6 subpath react-native-polyfill-globals requires.

useSearchContent — non-AI semantic search
Search & AI overview

Setup

Authentication

Users

Entities

Entity Lists

Comments

Reactions

Reputation

Spaces

Workspaces

Events

Chat

Collections

Relationships

App Notifications

Push Notifications

Search & AI

Storage

Moderation

Custom Tables

Usage Example

Parameters (`ask` function)

Returns

​Usage Example

​Parameters (ask function)

​Returns

​Related

Usage Example

Parameters (`ask` function)

Returns

Related