Seedlabs

Voice-to-Prompt Refiner

A tool that turns rambling voice notes into precise, well-structured prompts for AI.

A voice-first tool that turns a user's rambling spoken description of what they want into a clear, constraint-rich ChatGPT prompt, applying documented prompt-engineering best practices automatically — so people who struggle to write precise prompts can just talk.

Computer ScienceTopic Modeling
Accessibility and voice-productivity software

Concept

Most people know what they want from ChatGPT but cannot write it as a precise prompt. This tool lets a user speak their goal for a couple of minutes — tangents, disfluencies, and half-formed ideas included. It transcribes the speech, uses LLM semantic analysis to extract the underlying intent, entities, and constraints (the approach Orality [5] demonstrates for turning messy spoken content into structured representations), and then applies the prompt-crafting rules from the new guide [0] — clarity, explicit constraints, question-type fit, temperature setting — to output a ready-to-paste engineered prompt. Where the spoken brief is ambiguous, it asks one or two spoken follow-up questions before finalizing, mirroring the guide's guidance on handling ambiguous inputs. The wedge: prompt engineering for people who will never write like a prompt engineer.

Why now

The new guide [0] establishes that output quality hinges on deliberate prompt construction — clarity, explicit constraints, and ambiguity handling — which is precisely the skill most users lack, and it supplies the documented rulebook that the refinement step automates. Orality [5] demonstrates that LLM-based semantic analysis can convert disfluent, disorganized spoken content into structured, usable output, showing the speech-to-structure step is feasible today. Together the two papers cover both halves of the product: capture intent by voice, engineer the prompt by rule.

Written by GLM · 9 Oct 2026

AI assessment

Backed by 2 papers77

A practical utility that bridges the gap between messy human speech and structured prompt engineering, though it faces significant competition from native LLM voice modes.

Evidence strength
4/5
The idea effectively synthesizes a prompt-engineering guide with the 'Orality' paper's demonstrated ability to extract structure from disfluent speech.
Market pull
3/5
While there is a clear user pain point for non-technical users, the 'buyers' listed are mostly accessibility orgs or tools that may view this as a feature rather than a standalone product.
Novelty & moat
3/5
It is a distinct application of the 'Semantic Thought Mapper' logic, shifting the output from a visual map to a functional prompt.
Feasibility
5/5
The pipeline (STT -> LLM extraction -> Prompt Template -> Output) uses existing, mature APIs and requires no new fundamental research.
Wedge clarity
4/5
The focus on 'people who will never write like a prompt engineer' provides a sharp, specific user segment to target.
Simplicity / focus
5/5
The product is a single, focused utility with one clear input and one clear output.

Scored by Gemma 4 · 9 Oct 2026. Rated against a fixed rubric (evidence, market, novelty, feasibility, wedge, simplicity), relative to the closest ideas already in the catalog. A prior estimate to compare ideas before real-world signal arrives.

Persona discussion

AI personas trained on real people's expertise debate this idea as it evolves.

View the discussion →

Act on this idea

Ideas only matter if someone runs with them. Your message goes straight to the founder's inbox — nothing is stored on our servers.

Who benefits

  • British Dyslexia Associationorganization

    Its community often finds precise, structured writing difficult; speaking a goal and receiving an engineered prompt removes the writing barrier to using AI tools effectively.

  • Otter.aicompany

    Already transcribes what people say in meetings and notes; turning rambling spoken intent into a usable prompt is a natural next feature for its existing user base.

  • Grammarlycompany

    Helps people communicate clearly in writing and has moved into AI assistance; a prompt-refinement layer (typed or spoken) fits its product direction of improving what users send to AI systems.

  • Voiceittcompany

    Builds accessibility speech-recognition technology; a speech-to-prompt capability extends dictation from mere transcription to task completion for users with non-standard speech.

Research it builds on

  1. Prompt Engineering for ChatGPT: A Quick Guide to Techniques, Tips, and Best Practices
    Sabit Ekin · 2026 · 257 citations
    All ideas from this paper →
  2. Orality: A Semantic Canvas for Externalizing and Clarifying Thoughts with Speech
    Wengxi Li, Jingze Tian, Can Liu · 2026 · 2 citations
    All ideas from this paper →
Leave feedback
feasibility