gemini-share-readerlisted

# Gemini Share Reader ## Overview Extract public Gemini share pages by rendering the page with Playwright and collecting user/assistant messages in DOM order. Output clean conversation data as Markdown, JSON, or plain text for downstream summarization and analysis tasks. ## Quick Start 1. Validate the URL starts with `https://gemini.google.com/share/` or `https://g.co/gemini/share/`. 2. Run the extractor script: - `python3 scripts/read_gemini_share.py "<url>"` 3. Use output format as needed: - Markdown (default): `--format markdown` - JSON: `--format json` - Plain text: `--format txt` 4. Save to file when needed: - `--output /tmp/gemini_share.md` ## Workflow 1. Prefer the script in `scripts/read_gemini_share.py` for deterministic extraction. 2. Pass one share URL at a time; aggregate multiple links in a caller loop. 3. Prefer `--format json` when another tool will parse the output. 4. Prefer `--format markdown` when the next step is human reading or summarization. 5. If extraction returns zero messages, retry once with: - `--headful --wait-ms 12000` 6. If Playwright is missing, install dependencies and rerun: - `python3 -m pip install playwright` - `python3 -m playwright install chromium` ## Output Contract - JSON output structure: - `source_url`: input share URL - `final_url`: resolved URL after redirects - `title`: page title - `message_count`: extracted message count - `messages`: ordered list of `{role, content}` - Roles: - `use