Flomo Memo To Markdown

Overview

Convert flomo memos into Markdown files grouped by time range and generate tag-stats.md with both direct_count and total_count per tag node.

This is a standalone export skill. It is not a CRUD helper and should not be used for interactive "find one memo and edit it" requests.

Default behavior is optimized for fast one-shot processing from the local flomo desktop login state: full date range, monthly split, {{user}}_{{range}} filenames, and asset-mode=placeholder.

Preconditions

• flomo.app has been logged in on this Mac before
• Local flomo storage exists under ~/Library/Containers/com.flomoapp.m/...
• Network access to flomo API is available if using the local source

Quick Start

Set the script path from the current environment instead of hardcoding a user directory:

SKILL_ROOT="${CODEX_HOME:-$HOME/.codex}/skills/flomo-memo-to-markdown"
SCRIPT="$SKILL_ROOT/scripts/flomo_to_nblm.py"

Examples:

python3 "$SCRIPT"
python3 "$SCRIPT" --preview-only
python3 "$SCRIPT" --split quarter --asset-mode copy

Default Agent Workflow

1. Use the local flomo API source.
2. If any local API read step fails, fail fast and return error immediately.
3. Do not fall back to cached exports.
4. Do not ask for exported HTML unless the user explicitly requests postprocessing an archive.
5. Run with defaults unless the user specifies date range / split / naming / asset handling.
6. Report output directory, generated Markdown file count, per-file memo/char counts, threshold warnings, and the tag-stats.md path.

When To Ask Follow-Up Questions

Ask only when one of these materially changes execution:

• The user wants a custom date range
• The user wants a different split granularity
• The user wants a custom filename template
• The user wants copied attachments or explicitly wants to ignore attachments
• Local desktop auth is missing or broken

Output Behavior

• Default output is decompressed Markdown files
• Also writes tag-stats.md in export mode
• --preview-only performs parsing, filtering, grouping, and tag stats only
• --word-limit is treated as a character count threshold

Local Source Semantics

• Default input is the local flomo desktop login state, not exported HTML
• The script reads the local desktop auth token from ~/Library/Containers/com.flomoapp.m/.../leveldb
• Memo bodies come from flomo API content HTML and are rendered to Markdown
• Memo timestamps come from created_at
• Attachments come from memo files
• Tags are re-derived from rendered Markdown text using #tag syntax

Attachment Modes

• placeholder (default): Append attachment path lines to memo Markdown; do not copy files
• ignore: Ignore attachments entirely
• copy: Download signed attachment URLs into assets/ under the output directory and rewrite Markdown references

Tag Statistics Semantics

• Tags are extracted from memo plain text using #tag syntax
• Counting is deduped per memo
• direct_count: memos tagged with this exact tag
• total_count: direct_count plus all descendant tags

Resources

scripts/

• scripts/flomo_to_nblm.py: Main converter CLI for local flomo export to grouped Markdown

references/

• references/flomo-export-dom.md: HTML export DOM assumptions and attachment selectors