I built a platform for one-on-one language practice sessions

Regular practice is essential to mastering a language beyond A2. Teachers are expensive. I built a platform where speaking clubs can manage sessions through a shared calendar.

Solo Founder / Builder (2025):

  1. Built the platform solo using Cursor, Supabase, OpenAI API, Replit, Firebase, Telegram, and Brevo.

  2. Turned it into B2B SaaS for language schools and career agencies.

  3. Acquired B2B customers through cold outreach.

  4. Integrated an LLM API, doubling the platform's audience.

Create a speaking club where members book 1:1 sessions

Book 1:1 sessions with your club's members

Key facts

Key facts

Why

  1. Millions get stuck at B1, lacking fluency to speak English.

  2. Learning apps teach the rules, not real-life communication.

  3. Private tutors charge $40+ per online session.

  4. No efficient way existed to find regular speaking partners with shared interests.

Business

  1. SaaS for language schools and career agencies to engage and retain clients. $99/month.

  2. MVP growth: target existing communities individually to overcome the cold-start problem.

  3. Acquired first customers through targeted cold emails.

Product

  1. Closed group calendars for booking 1:1 sessions.

  2. Google authentication with ToU & PP acceptance.

  3. Scheduling: sharing slots, requesting, booking, and cancelling sessions.

  4. Email and Telegram notifications.

  5. Private chat for rescheduling without sharing contact details.

  6. User profiles with languages, interests, LinkedIn, and notification settings.

  7. P2P mock interviews with AI-generated personalised questions.

Share recurring slots or set specific availability

Manage incoming requests

Get notifications and chat via email or Telegram

Mock Interview Practice: AI-Powered Feature

Mock Interview Practice: AI-Powered Feature

Users create tailored interview questions based on their CV and job description to practice mock interviews.

Practice P2P mock interviews with questions tailored to your CV

How I built it

How I built it

After building a demo-ready MVP, I found a design partner to co-develop the service. They became my first client.

Typical feature development process:

  1. Explored new technical topics (e.g., auth tokens, cron jobs) with ChatGPT.

  2. Detailed the implementation pipeline for Claude in Cursor, down to individual queries, as LLMs still struggled with autonomous development in mid-2025.

Typical feature development process:

  1. Explored new technical topics (e.g., auth tokens, cron jobs) with ChatGPT.

  2. Detailed the implementation pipeline for Claude in Cursor, down to individual queries, as LLMs still struggled with autonomous development in mid-2025.

Some of the most frequently used pipelines on the platform:

Some of the most frequently used pipelines on the platform:

Outcomes

  1. Alla helped two communities of 350+ members engage in P2P language practice.

  2. Hundreds of conversation hours spent sharing career advice, life stories, mutual support, and practising languages.

  3. Improved the product through user interviews and data, reaching 12% week-2 retention.

Outcomes

  1. Alla helped two communities of 350+ members engage in P2P language practice.

  2. Hundreds of conversation hours spent sharing career advice, life stories, mutual support, and practising languages.

  3. Improved the product through user interviews and data, reaching 12% week-2 retention.

Outcomes

  1. Alla helped two communities of 350+ members engage in P2P language practice.

  2. Hundreds of conversation hours spent sharing career advice, life stories, mutual support, and practising languages.

  3. Improved the product through user interviews and data, reaching 12% week-2 retention.

Top members' meetup. Yaroslav shares how he just landed a job abroad

© Kostya Puliarkin, 2026

© Kostya Puliarkin, 2026

© Kostya Puliarkin, 2026