Postery: autoposting to Telegram and VK with AI

Automating Publications with Postery: From Creation to AI Generation

Developers and content creators often spend hours adapting posts for different platforms: Telegram channels, VK groups. Postery solves this problem through a unified interface. The tool allows you to create a post once, customize variations for sources, and schedule publication. Support for an unlimited number of channels of the same type simplifies cross-posting without duplicating effort.

The process begins with adding sources. For Telegram, channels or groups are connected via a bot token; for VK, via an access token. After verification, sources appear in the list, ready for use.

Creating and Adapting Content

At the editing stage, the post's title, description, tags, and images are specified. Next, target sources are selected. The key step is customization: for each source, the text and title are edited considering platform limits. Telegram requires brevity (up to 4096 characters), while VK allows detailed posts with hashtags.

• Benefits of Adaptation:

- Natural appearance of the post on each platform

- Compliance with length and formatting limits

- Addition of platform-specific elements (emojis, links)

This feature minimizes manual work while maintaining the content's core meaning.

AI Integration for Text Generation

Postery integrates AI providers GigaChat and OpenAI to automate content. In source settings, prompts are set: style (business, creative), structure (introduction + list + call to action), character limits.

When creating a post, the generate button triggers AI: for Telegram—a concise version; for VK—an expanded version with keywords. Auto-mode generates texts for all sources at once.

Prompt Setup Example:

"Generate a title up to 100 characters and a description up to 2000 in [style] style. Structure: hook + 3 facts + CTA. Topic: [post topic]."

AI ensures style consistency across platforms, reducing copywriting time by 70–80%.

Scheduler and Analytics

Each post is assigned a date/time for publication. Multi-scheduling is supported: one post goes out on Telegram today, on VK—in a week. The calendar visualizes the schedule in a day grid with time details.

The analytics section tracks:

• Published posts
• Queue for publication
• Sources and statuses

Automatic publication works on a cron-like schedule, without manual intervention.

Technical Implementation with FastAPI

Postery is built on Python with FastAPI for the backend. The architecture is stateless, with PostgreSQL for storing posts/sources and Redis for task queues (Celery). Local launch: pip install -r requirements.txt && uvicorn main:app.

Frontend on Vue.js with Tailwind CSS. API endpoints:

• /posts/ — CRUD for posts
• /sources/ — source management
• /schedule/ — scheduling
• /ai/generate/ — LLM calls

Docker-compose for deployment: containers for app, db, redis, celery. Configuration via .env: API keys, DATABASE_URL.

The project is open source, at MVP stage: basic functionality is stable; roadmap includes Mastodon, Twitter/X, email newsletters.

Key Takeaways

• Postery automates 90% of publication routine through a single dashboard.
• AI integration (GigaChat/OpenAI) generates adapted content based on prompts.
• Multi-scheduling with a calendar for staggered releases.
• FastAPI backend is scalable, ready for self-hosting.
• Open source on GitHub, community influences development.

— Editorial Team