Any chat.

You do not need a particular app’s .txt. Ask any AI to guide the export and convert it to Left on Read’s canonical JSON. The app parses the JSON on your phone.

The prompt

I want to import a chat into Left on Read, a private, local-first AI mirror app. Help me export my chat from <APP> on <iPhone / Android / desktop>.

First, tell me the exact steps to export that chat. Then, if I paste the export, convert it into Left on Read's canonical JSON. Do not invent messages. Keep the original senders, timestamps and order. Mark media, deleted messages and system events. Output only valid JSON.

Schema:
{
  "format": "left-on-read.import.v1",
  "platform": "telegram",
  "me": "my display name",
  "them": "the person I want to mirror",
  "participants": ["my display name", "their display name"],
  "messages": [
    {"ts": "2024-01-01T12:00:00Z", "sender": "my display name", "text": "hi"},
    {"ts": "2024-01-01T12:01:00Z", "sender": "their display name", "text": "haan"},
    {"ts": "2024-01-01T12:02:00Z", "sender": "their display name", "kind": "media", "text": "<image omitted>", "mediaType": "image"},
    {"ts": "2024-01-01T12:03:00Z", "text": "Nisha joined", "kind": "system"}
  ]
}

Rules:
- ts must be ISO-8601 when possible; keep the original wall-clock time.
- kind is text (default), media, deleted or system.
- System events have no sender.
- Do not summarise, rewrite or merge messages.
- If the export is HTML, JSON or another shape, map it into the schema; do not drop lines.
- Return only JSON, with no markdown fences or commentary.

Privacy: I understand pasting a chat into an AI is my choice. Left on Read itself parses the final file on my phone and never uploads the archive.

Only paste a chat into an AI you trust. Left on Read never uploads the archive; the converted JSON is parsed on your device.