Skip to content

Releases

v1.0.0 — DocuMentor AI (Go CLI, Zero-GEMINI_API_KEY Host-AI MCP Server & 7-Client Skill)

Choose a tag to compare

@tanaikech tanaikech released this 08 Nov 03:28

gdocumentor v1.0.0 — Initial Release

gdocumentor (DocuMentor AI Go CLI, Dual-Engine MCP Server gdocumentor mcp, and Multi-Agent Skill / Plugin) turns Gemini and any MCP-compatible AI agent (Google Antigravity agy, Claude Desktop, Claude Code, Cursor, VS Code / GitHub Copilot, Windsurf, and Gemini CLI) into a Socratic writing & code review coach and 0-token Google Docs automation engine.

Highlights

  • Dual AI Execution Engine: Run as a standalone Go CLI via GEMINI_API_KEY (gemini-3.8-flash / gemini-4-argon), or as an MCP Server (gdocumentor mcp) powered by your client AI agent's own LLM (Host AI, 0 GEMINI_API_KEY required).
  • Zero-Drift Bottom-Up UTF-16 Inline Suggestions & Socratic Hints: Applies WriteMode.SUGGEST inline diffs with 💡 Reason and 🎓 Learning Point replies, or withholds answers in Socratic Hint Mode (--hint) so human authors write the fix.
  • Multimodal Prose × Table (TABLE_1) × Figure (VISUAL_1) Verification: Cross-checks narrative claims against embedded images and tables.
  • Scoped Code & Manuscript Review (--target, --auto-deps): Restricts edits strictly to specified line/character ranges (#L1-L35, @0:2500) while automatically loading imported files as read-only context.
  • AI (🤖 AI) + Human (👤 Human) Review Learning Summary (gdocumentor summary) & 4-State Decision Record: Analyzes both AI-generated and human-authored suggestions/comments on existing Google Docs to synthesize personalized learning advice, and tracks Accepted, Edited, Rejected, and Expired outcomes.
  • Zero-Token Direct Google Docs Editor (gdocumentor edit): Creates Pageless Google Docs, converts/inserts rich Markdown + tables + local images, and resizes VISUAL_1 in place with 0 LLM tokens.
  • Self-Provisioning 7-Client Installer (gdocumentor setup) & 1-Command WSL2 ↔ Windows 11 Bridge (--wsl-host): Non-destructively installs MCP server configs and native agent skills/rules across all 7 AI clients, including automatic wsl.exe --exec cross-OS bridging.