Friday Transcript

Conversation records, SQLite/JSONL stores, and CSV, JSONL and Markdown exporters. Use transcripts for session history and inspection; use Friday Memory for extraction, semantic recall and context assembly.

Distribution: friday-framework-transcript · Import: friday_transcript Python: 3.13.4+ · Status: alpha · License: MIT

Installation

python -m pip install "friday-framework-transcript==0.1.0a4"

Friday distributions use the same release version; keep internal Friday dependencies aligned.

Quick start: persist a conversation

import asyncio
from pathlib import Path
from tempfile import TemporaryDirectory

from friday_transcript import SQLiteTranscriptStore, TranscriptEntry, TranscriptSession

async def main() -> None:
    with TemporaryDirectory() as directory:
        store = SQLiteTranscriptStore(Path(directory) / "transcript.db")
        await store.create_session(
            TranscriptSession(session_id="demo", model_name="offline-example")
        )
        await store.append_entry(
            TranscriptEntry(
                id="entry-1", session_id="demo", turn_number=1,
                user_message="Hello", assistant_message="Hello from Friday",
            )
        )
        entries = await store.get_entries("demo")
        print(entries[0].assistant_message)

asyncio.run(main())

Expected output: Hello from Friday. The temporary database is removed when the example exits. Supply an application-owned persistent path to retain it.

Records and storage

• TranscriptSession identifies a session and its configuration/aggregates.
• TranscriptEntry records a turn, including optional tool, usage, timing and provenance fields.
• SQLiteTranscriptStore supports querying and snapshot/recovery workflows; JSONLTranscriptStore provides a file-oriented alternative.
• Exporters create portable views. Exporting records is separate from semantic-memory ingestion.

Transcripts can contain prompts, responses, tool results and configuration snapshots. They are not covered by content-free telemetry guarantees. Apply your own retention, access and redaction policy. Coordinate maintenance through the public quiescence/snapshot contracts rather than copying or replacing a live database.

Development

Work from the monorepo root after the shared contributor setup. There is no package-local test suite yet. Start with an import check:

uv run --no-sync python -c "import friday_transcript"

SQLite behavior is also exercised by Runtime and Chat integration tests. Run the relevant consumer suite when changing persisted records or store contracts.

Documentation and support

Memory integration · Runtime integration · Source · Issues · Contributing

License

MIT.