Products and tools I build, operate, and use in real workflows.
whytap
live product · macOS and WindowsVoice input and meeting intelligence for individuals and teams.
Whytap provides fast voice input, automatic meeting notes, terminology-aware speech recognition, and search across personal and shared meetings. Meeting knowledge can be exposed to internal AI assistants through MCP.
Open live app ↗voice AI · meeting intelligence · MCP · speech-to-text
session recall
open source · in daily useShared memory for Claude Code, Codex, and Cursor.
Session Recall builds one searchable index across coding-agent sessions, allowing one engine to retrieve decisions, architectural trade-offs, bugs, and fixes discovered by another. Semantic search is exposed through MCP, while raw turns, tool traces, and reasoning remain available on demand.
View on GitHub ↗shared agent memory · semantic search · MCP · SQLite
storm
in daily useA multi-model software engineering workflow that runs codebase analysis, planning, implementation, review, and debugging through several coding agents in parallel.
Instead of trusting a single answer, Storm compares independent solutions and consolidates them through an explicit review stage.
multi-agent systems · coding agents · evaluation
langlearn
live · langlearn.maxbutorin.devA voice AI tutor for English: talk to an agent, get a cleaned bilingual transcript with tap-to-translate, then a personal lesson generated from the conversation.
Open live app ↗voice AI · Next.js · ElevenLabs · OpenRouter