AI / PERSONAL KNOWLEDGE / TELEGRAM

A second brain that turns voice and messages into connected memory

VoiceMind is a personal AI assistant inside Telegram. It accepts voice, text and documents, preserves the original context, turns a stream of thoughts into notes, tasks, ideas and projects, and then helps the user work with accumulated memory through natural conversation.

  • PHP
  • Telegram Bot API
  • SQLite
  • FTS5
  • OpenAI-compatible API
  • TickTick API
  • Cron
  • Webhooks
VoiceMind — project preview

Why VoiceMind was created

Why VoiceMind was created

Important thoughts, decisions, promises and project details often remain trapped in voice messages and scattered notes. VoiceMind provides one place where the user can speak or write without preparing a structure, while the system preserves the source, identifies the record type and brings the right context back when needed.

Who the system serves

The product supports people who run several projects, capture ideas, make many decisions and want to preserve a personal history without constantly organizing notes by hand. Telegram remains a familiar conversational interface while structure is created inside the system.

01

Challenge

Turn an unrestricted stream of voice and text into long-term memory without losing the original record, event date, project context or relationships between entities. AI processing must not delay the Telegram webhook, and repeated delivery of the same event must not create duplicates.

02

Solution

We implemented a PHP Telegram webhook that stores the original input before placing processing into a queue. A background worker transcribes voice, detects intent, structures the record and updates memory. SQLite stores sources, records, projects, conversations and summaries, FTS5 and a hierarchical index retrieve relevant context, and TickTick connects tasks to external planning.

Capabilities

Product capabilities

The capabilities support one operating scenario and share common state, control and analytics.

01

Voice and text capture

The assistant accepts regular messages, voice notes, audio files and text documents while preserving the original material before AI processing.

02

Structured memory

The incoming stream becomes notes, ideas, tasks, reminders and journal entries with dates, tags, priority and project relationships.

03

Contextual conversation

Users ask questions in natural language while the system retrieves recent records, full-text search results, summaries and its compact memory index.

04

Projects and tasks

VoiceMind creates and updates projects, tracks their related tasks and supports task creation and completion through TickTick.

05

Daily summaries

A scheduled process collects the day's records, identifies key events and open loops, and stores the result together with source references.

06

Long-history context

The system retains conversation context, autobiographical records, event dates and periodic summaries so work can continue without losing previous decisions.

Outcome

Outcome

VoiceMind turns Telegram from a channel of scattered messages into a personal long-term memory workspace. Voice and text follow one path from original source to structured record, while search, conversation, daily summaries and tasks use the accumulated context as one connected system.

PHP Telegram Bot API SQLite FTS5 OpenAI-compatible API TickTick API Cron Webhooks

New project

Working on a similar task?

Describe the current situation and the outcome you need. We will understand the context and propose a practical first stage.