logo

MindSynth Alpha

Co-create custom interactive systems with a network of agents powered by language models. Easily spin-up novel applications, and interact with existing apps to reconfigure them around your unique creative workflow.

Design View

Roadmap

Foundational Research

Develop core methodology and techniques for generating interactive apps based on cognitive representations

Prototype Platform Architecture

Technical feasibility research and development of initial prototypes

Prototype Persistent Data Store

Initial implementations including in-memory graph, json graph and Neo4j persistent store

Native Audio Graph Initial Implementation

Prototyping with macOS audio engine API for basic audio functionality

POC for Generating Native App

Generate functional app state from data structure passed in at runtime

User Research 1

Evaluate concepts and collect insights into user tasks, goals, and desired functionality

3D Representations Concept Development

Develop visual and interactive concepts for audio data interaction in 3D space

3D Interaction Web Demo

Demo implementation for working with a timeline and audio data objects in a 3D context

Foundation Model Integration 1

Initial research and experiments using foundation model results in program synthesis pipeline

Agent Implementation 1

Multi-agent logic for mediating between user and model and coordinating task-specific agent processes

Web Demo Concept Development

Audio data interaction concepts across data processing contexts

Web Demo UI

UI design integrating core front-end input and visualisation modules

Web Demo Audio Graph POC

POC audio graph implementation to support real-time interactions

Native Demo Audio Graph POC

Updated version of audio graph for macOS prototype

Agent Implementation 2

Updated architecture for in-context learning, planning and flexible query response

UI Design 2

Iteration on DesignView and initial DomainView design

Database server implementation for PersistentMemory module

Read and write logic to support each agent's persistent memory spec

User Research 2

Alpha prototype research

Open Source Model Evaluation

Evaluate efficacy of structured outputs across agent tasks

Sketch Interaction Implementation

Add sketch interaction functionality to web demo result view

Vision API Integration

Add support for vision API in program synthesis pipeline

DB Service Integration

Integrate database to support persistent representations across sessions

Agent Implementation 3

Iterate on multi-agent system to support broader set of user design specifications

Web Demo Production Build

Infrastructure and updated functionality to support production version of demo

Foundation Model Integration 2

Updated logic and issue fixes given web demo feedback / finalise model for Beta

UI Design 3

Beta version UI

Beta Production Build

Develop production version for Beta launch

User Research 3

Beta programme user insights collection and analysis

Agent Implementation 4

Updated multi-agent system development given final target specification

Foundation Model Integration 3

Model integration for finalised pipeline design

UI Design 4

Finalised design for launch version

Production Build

Functionality and optimisations for launch version

User Research 4

Production app user insights collection and analysis

Get notified on beta access

Subscribe for updates on early access, product launches and company news

Made in London