Mouth Coding: AI-Facilitated Collaborative Web Building
Mouth coding uses real-time conversations with LLMs, transcription, and live previews to build websites collaboratively, prioritizing human judgment to create inclusive designs faster—ideal for small teams and non-profits.
Essential Ingredients for Mouth Coding Sessions
Build websites by conversing with an LLM: start with a group discussion on values, mission, and goals, then refine a shared specification before generation. Key components include live speech-to-text transcription (e.g., Notion AI for seamless note integration), sturdy UI infrastructure like design systems to ensure quality outputs, and real-time previews in tools such as Storybook or experimental UI environments. Pull in additional context like existing sites, design explorations, and best practices during the talk. Human elements—taste, critical thinking, and skills—guide decisions, preventing poor results. In one hour, this produced a vastly improved site for Lucent Counseling by iterating verbally on layout, content scaffolding, and mission alignment, yielding a prototype ready for tweaks like changing headings.
Trade-offs: Avoid free-reign generation without infrastructure, as it risks low quality. Sessions don't finish everything—final copy must be human-written (AI only for scaffolds), plus deploy via Netlify and domain transfers—but these steps are now far less laborious.
AI as Collaboration Facilitator, Not Replacement
Mouth coding counters siloed workflows (e.g., Jira tickets, designer-dev toggles) by enabling genuine cross-disciplinary input: fewer oversights, on-demand context retrieval, surfaced blind spots, and clearer next steps. Participants see changes instantly (e.g., 'Make the heading say “Better together”'), treating the site as iterable clay for ideation and refinement until it feels right or time ends. This democratizes design, including non-experts like therapists or small business owners, whose mission clarity shines without deep coding skills.
Core rule: AI facilitates human creativity—reiterated emphatically—leveraging collective perspectives for richer outcomes. Result: More human-feeling digital collaboration than in years, with prototypes that excite and educate users on iteration potential.
Impact on Non-Profits and Small Businesses
Target underfunded, mission-driven groups with outdated sites due to thin resources: owners articulate visions verbally, pros facilitate sessions pro bono more feasibly thanks to speed. Enables diverse voices to shape work advancing healing, trauma recovery, or community goals. Workshops incoming; author seeks non-profits for public demos to teach the method. Professionals gain efficient community service without full-time commitment, benefiting all.