ttmi makes your AI agent draw diagrams instead of dumping walls of text
blixt · x · 2026-10-20
Developer christianalfoni launched ttmi (time to mental image) to fix the "wall of text" problem with agent explanations: instead of returning twelve paragraphs, the agent draws what it means and walks through it step by step, narrated by in-browser voice that you can tune to pronounce your product names and jargon.
The layout engine claims to encode diagram-comprehension research (e.g. Larkin & Simon 1987's locality principle) as enforced rules rather than style suggestions. It works across code, infrastructure, business processes, contracts and more. Pricing: free for individuals (unlimited sketches), with a Team tier adding org-wide sharing, PNG export, CI API keys (a sketch on every pull request), and SSO.
More from coding & agent
- VibesPlug open-sources 147 cross-IDE coding agent skills with CLI and MCP server — Roger_M_Taylor · 2026-10-20
- Beacon raises $5.5M for open-source cross-tool memory layer for coding agents — HeyNayeem · 2026-10-20
- Beacon raises $5.5M seed for open-source cross-tool memory layer for coding agents — Scobleizer · 2026-10-20
- One sentence, 300 Kimi K3 agents, zero code: a 7-step agent build workflow — Roger_M_Taylor · 2026-10-20
- FrankenTerm plans air-gapped transfer of Mac settings and SSH keys via animated QR video — doodlestein · 2026-10-20
- How engineers rescued a vibe-coded food-safety app serving KFC and Taco Bell kitchens — alex_verem · 2026-10-20