* Accepts either: * - JSON: `[{ speaker, content, timestamp? }, ...]` * - Plain transcript: `Alice: Hello\nBob: Hi` (one message per line) * - WeChat-export style: `[2024-01-01 12:00:00] Alice\nHello`
(
raw: string,
chunkOpts: { targetTokens: number, overlapTokens: number },
filename?: string,
)
source not stored for this graph (policy: none)
no test coverage detected