Automation of event search and publication in Telegram and Facebook

Budget: 20000 UAH Deadline: 14 days

Good day, Artem! I was gathering exactly such pipelines: event source → AI processing and adaptation in Ukrainian → admin approval in the Telegram bot → queue → auto-publication. We will create an MVP turnkey.

Stack: I propose my own code in Python on our VPS instead of n8n/Make — a modular architecture where each event source is a separate adapter, so Finishers can easily be replaced or new ones added later without rebuilding the system. AI — OpenAI/Claude, publication via Telegram Bot API and Meta Graph API, queue and statuses in a lightweight database (SQLite/Postgres) or Google Sheets. My code is cheaper to maintain than subscriptions to builders, and we support it.

Experience: I manage several Telegram bots and content pipelines with AI processing and auto-posting — this is our specialty.

Finishers: first, I check the official API/export; if it’s not available — a robust parser with repeat logic and a backup source in case of limitations. In the publication, we insert the canonical link to the official website of the event/organizer, not to Finishers — we will establish this in the adapter right away.

Photos: we do not pull others' images without rights. Options for approval in the bot — official photo of the organizer when the license allows; selection from free banks (Unsplash/Pexels) based on the type of event; or a branded template card. The admin chooses in the preview.

MVP deadline: 12-14 days. Cost: 20,000 UAH.

One clarification: is there already a Facebook page and Meta business account for publications (Meta API requires a verified application/page), or should we set this up as well?

I am ready to show a live demo of the approval bot in the chat so you can see the flow to the solution.

Petro Pankov, BotCraft Group Expand

Budget: 20000 UAH Deadline: 10 days

The main risk here is not in the automation of publications itself - the combination of parsing + AI + bot + auto-posting has long been established. The risk lies in the fragility of the source: Finishers does not have an official API, so any automated data retrieval could break with changes in markup or the introduction of protection. The second slippery point is image rights during auto-publication. Both issues need to be incorporated into the architecture from day one, and that is exactly how I plan to proceed.

Regarding your points:

1. Stack: n8n (I prefer it for the flexibility of self-hosting, but I am open to Make.com), Claude or OpenAI for AI adaptation of texts in Ukrainian, Telegram Bot API for the approval bot, Meta Graph API for Facebook, queue and status of publications - Google Sheets/Airtable or a simple database (SQLite/Postgres).

2. Experience: I have been programming since 2013. In production, I have my own automation system: task scheduler + LLM content processing + delivery of results via Telegram Bot API - essentially the same combination you need. A public example of work with parsing and data cataloging is pinger.md (an ad aggregator).

3. Finishers: I will start with a technical audit (hidden API, RSS, sitemap, structured data in markup) - often such sites have more stable channels than HTML. Next, a parser with caching, retries, and monitoring of markup changes with admin alerts. I will isolate the source of events with an adapter: adding or replacing a source is a new module, not a system overhaul. Fallback: official calendars of organizers and alternative aggregators.

4. Photos: I propose a three-tier scheme. By default - auto-generation of branded covers according to a template (title, date, city, type of sport): no questions about rights and a unified style for the feed. If the organizer publishes official press materials - we use them. Fallback - stock images with a free license via Unsplash/Pexels API.

5. Deadline: 10 days.

6. Cost of MVP: 20,000 UAH, fixed.

Work plan: audit of Finishers and fixing the method of data retrieval, parser and normalization of events, AI adaptation in Ukrainian with a link to the official event website (not to Finishers), Telegram bot with previews and approval buttons, queue and auto-posting in Telegram and Facebook.

Two clarifications: how many new events do you expect per day, and is there already a Facebook page with access to Business Manager?

I am ready to discuss the details in correspondence. Expand

Budget: 20000 UAH Deadline: 12 days

Good day! I have done such turnkey connections — source parser, processing through OpenAI in Ukrainian, agreement in a Telegram bot, and auto-publication in TG and on the FB page via Meta API. For Finishers, I will immediately lay out a backup plan: there is no official API, so I will take the data with a careful parser with a fallback in case of restrictions. I will make the architecture modular on n8n, so that new sources can be added without reworking. Approximately 12 days, I will fit into your budget. I will look at your specification file and clarify the issue with photo rights.

Budget: 20000 UAH Deadline: 12 days

Artem, hello. The main risk here is not AI generation, but the stable acquisition of data from Finishers and the legal use of photographs. Therefore, I will implement Finishers as a separate module-source: if access or the structure of the site changes, it can be replaced or supplemented without rebuilding the entire system.
I already have a Python system that regularly retrieves new entries from an external API, checks them for duplicates, processes the data through OpenAI, stores the history, and sends the result to Telegram. For your MVP, I will use this proven architecture, adding content moderation, photographs, a queue, and publication through the Meta API.
Recommended stack: Python/FastAPI, PostgreSQL, OpenAI, aiogram, Telegram Bot API, and Meta Graph API. I will deploy the system on Railway or VPS in the client's account — the code, database, keys, and all accesses will remain with you. This will avoid the limits of Make/n8n and reduce monthly costs.
Before development, I will check the official mechanisms and the public structure of Finishers. If there is no stable API, I will use a controlled parser for open pages with caching, rate limiting, and an error log, without bypassing the site's protection. Duplicates will be identified by ID, date, and the canonical link of the event. Only confirmed links to the official website of the organizer or the registration page will be published; questionable events will go for manual verification.
For photographs, I propose a hybrid option: an automatic branded event card based on your or licensed backgrounds; an official promo photo, if its use is permitted; or manual replacement via a Telegram bot. The source of the photograph will be stored along with the event.
The process will work like this: event search → duplicate check → AI adaptation in Ukrainian → preview in Telegram → buttons "Approve," "Replace photo," "Reject" → queue for the nearest available slot → publication in Telegram and Facebook. The AI prompt and schedule can be changed without editing the code.
The implementation period is 12 calendar days. The final cost of the MVP is 20,000 UAH. Estimated monthly expenses are $10–25 depending on hosting, the number of events, and the volume of AI processing. The cost includes development, deployment, testing, code transfer, prompts, database structure, and a short instruction.
Please let me know if the Facebook page is already connected to Meta Business Manager and if there is administrative access for setting up official auto-publication. If needed, I can show a similar working monitoring system and Telegram bot.

Budget: 22000 UAH Deadline: 12 days

Good day, Artem. I read the description and the 4-page technical specification.

1) Stack: n8n + a small PostgreSQL/SQL database for states, queues, and deduplication, OpenAI, Telegram Bot API, official Meta Graph API; JS code only where n8n nodes are insufficient.
2) Relevant experience: I have a self-initiated demo (not a client case) n8n → OpenAI → human approval → Telegram, with retries, rate limits, and importable workflow: https://prooflab-live-cases.dolponik.chatgpt.site/demos/automation
3) Finishers: I will start with a short feasibility spike — I will check allowed public structured endpoints/HTML, field stability, and access rules. No bypassing protection. If there are limitations, I will suggest an alternative source or the official pages of the organizers.
4) Risks: markup changes, incomplete/conflicting data, rate limits/ToS; I will cover validation, cache, retries, and alerts.
5) Duplicates: source ID/URL + normalized event signature (name/date/city), unique key, status history.
6) Photos: for MVP, I recommend a safe semi-auto flow — photos only from allowed/official sources with provenance tracking, preview, and add/replace in Telegram. I do not use random image scraping.
7) Telegram: preview + inline Approve/Reject + optional photo replacement; only approved entries go into the queue.
8) Publication: scheduled_at/status/retry_count, Telegram and then Meta through the official mechanism, idempotency, and error log.
9) Monthly: preliminarily $10–35 for a moderate volume (VPS + actual AI consumption; the database can remain on the free tier). More accurately — after the volume of events and accesses.
10) Deadline 12 days, final cost 22,000 UAH. Includes setup in the client's accounts, test run, transfer of workflow/code/prompts/database structure, and a short instruction.

I will start by checking the Finishers and Meta accesses, after which I will finalize the MVP architecture. Expand

Budget: 20000 UAH Deadline: 10 days

Good day!

I can implement such an MVP. I suggest obtaining data through parsing (or API, if available), then processing it using AI, sending it for approval via a Telegram bot, and automatically publishing it on Telegram and Facebook.

I have experience with Laravel, automations, Telegram Bot API, and I have already implemented a similar system for data collection, AI processing, and publication.

For photos, I will propose a secure solution considering copyright. After reviewing the technical specifications, I will be able to suggest the optimal stack, estimate the timeline, and provide the final cost.

Budget: 20000 UAH Deadline: 10 days

Hello! I'll go through your list.

Stack: n8n as an orchestrator (it allows adding new sources without rebuilding, as you want) plus OpenAI or Claude for text adaptation, Telegram Bot API, Meta Graph API, and a simple database (Postgres or Airtable) for event history. All accounts and keys are yours.

Finishers: first, I will check if the front pulls events through a hidden JSON request; it's more stable to hook into than parsing HTML. If not, I will use headless with a human pace. The risk is real: the site may change its layout or restrict access, so I will create sources through adapters to replace Finishers or add others without reworking everything. To avoid duplicates: I will remember each event by id/link and not process it again. The final link in the post will lead to the official event website, not to Finishers.

AI selects facts, translates, and adapts them in Ukrainian with a lively style, without making anything up beyond the source. I will carry the prompt in the settings to change it without me.

About photos separately, as requested. Scraping someone else's photos from Finishers is not worth it; those are someone else's rights. I will take the main photo from the official page of the organizer (their promo); if not available, I will substitute from a small library of licensed images based on the type of activity (running, cycling, trail). And I will leave the admin the option to replace the photo directly in the Telegram bot. This way, it is both automatic and clean in terms of rights.

Approval: the bot sends a preview (photo, text, link) with Approve/Reject buttons, and photo replacement is also done through the bot. Approved items go into the queue for the nearest available slot according to a customizable schedule: Telegram via Bot API, Facebook to your page via the official Graph API, this is an allowed path without the risk of bans.

Monthly approximately $15-40: AI tokens for this volume are not much, n8n hosting, and optionally a paid image bank. The term is 10-14 days. The final cost of the MVP is within your budget, $550. I would be happy to collaborate. Expand

Budget: 20000 UAH Deadline: 9 days

Good day!

For the MVP, I propose the stack: n8n, Supabase/PostgreSQL, OpenAI API, Telegram Bot API, and Meta Graph API.

I have experience with n8n, AI data processing, Telegram bots, Google Sheets, and API integrations.

Before development, I will check how to reliably obtain data from Finishers: through available structured data, website queries, or other permissible methods. The source module will be separate, so Finishers can be replaced in the future without rebuilding the entire system.

I will implement duplicate protection through the event ID or a unique combination of name, date, city, and official link.

For photos, I propose a semi-automatic option: the system takes images from the official event website, the administrator sees it in Telegram and can approve or replace it. If there is no photo, the system creates a branded card with the name, date, and city.

After approval, the post goes into the queue and is automatically published on Telegram and Facebook. There will also be a status log and the ability to change the schedule and AI prompt without a developer.

Timeline: 10–14 days.
Cost of MVP: $550.
Estimated monthly expenses: $10–30.

After completion, I will provide the scripts, database, prompts, settings, and a brief instruction. Expand

Budget: 19000 UAH Deadline: 9 days

Igor, good afternoon.
I have reviewed the task. Such automations are quite familiar to me and have been done before: data parsing, processing through AI, publication approval in Telegram, and automatic posting in Telegram and Facebook.
I understand that the main point here is not the publication itself, but the stability of data retrieval from Finishers. Therefore, I will make the system modular so that in the future it will be easy to replace or add new sources if the website changes its structure.
I mainly work with n8n, integrations through API, and AI services, so I will be able to implement the entire process turnkey, from event retrieval to the final publication after confirmation in Telegram.
Before starting, I will analyze the technical specifications in detail, and if necessary, I will ask you more questions to immediately consider all nuances and avoid rework during the process.
I would be happy to discuss the details and answer any questions, feel free to message me if you have any.)

Budget: 20000 UAH Deadline: 7 days

Good day.

Ready to assemble the MVP: Finishers → AI (Ukrainian) → approval in Telegram → queue → publication in Telegram and Facebook with a link to the official event/registration site, not to Finishers. Architecture — modular "sources," to later add other sites without rewriting the core.

— Stack
n8n (orchestration) + PostgreSQL or Airtable (queue/statuses) + OpenAI (text adaptation) + Telegram Bot API (preview/approve/reject) + Meta Graph API (Facebook). Finishers parser — a separate lightweight Python service or HTTP node in n8n with adapter source=finishers. No unnecessary custom backend — cheaper to maintain.

— Experience
Autofunnels n8n/Make, Telegram bots with approve flow, content generation via LLM, social media publication through API, post queues with statuses draft → approved → published.

— Data retrieval from Finishers (honestly and reliably)
Initially, a 0.5–1 day audit: whether there is RSS/embed/public JSON, URL structure of events, how often the feed is updated, what about robots/ToS.
Optimal: official/semi-structured feed, if available.
If only the website — cautious HTML parsing of public cards (interval, cache, idempotency by event ID), without bypassing captcha/antibot.
If access is restricted: (1) manual CSV/Google Sheet as a fallback source; (2) alternative calendars/sites of organizers with the same adapter; (3) webhook "added event manually in Sheet → then the same AI+approve pipeline." In the MVP, I include a source field and a normalized event model (name, date, city, official registration URL, description).

— Photos
I do not pull images from Finishers "as is" — risk of copyright. Options for MVP:

Official event photo, if the organizer provides an open OG-image / press kit and the license allows (best option).
Otherwise — neutral template (own branded covers: running / triathlon / swimming) + text; the link leads to the official registration.
Optionally later: Unsplash/Pexels by keywords (only licensed stock) or photo upload by admin with a button in the bot upon approval.
In the post, the priority is always the official URL, not Finishers. Expand

Budget: 20000 UAH Deadline: 14 days

Good day!

The task is close to me: I am currently developing my own AI assistant with a similar architecture — data retrieval via API/scraping, AI analysis, draft creation, user confirmation, action execution, logging, and protection against duplicates. This is not a ready-made copy of your system, but its technical core is very similar.

For the MVP, I propose the following stack:

— self-hosted n8n for scenarios and scheduling;
— a small Node.js module for working with Finishers;
— PostgreSQL for events, queues, statuses, and history;
— OpenAI API for translation and text adaptation;
— Telegram Bot API for approval;
— Meta Graph API for publishing on a Facebook page.

Before the main implementation, I will conduct a brief technical check of Finishers: the presence of an official API, feed or structured data, rules for automated access, and page stability. I do not propose building a system on private or hidden APIs.

If scraping public pages is allowed, I will use a limited request frequency, caching, retries, and structure change control. The source will be a separate adapter, so Finishers can later be replaced or supplemented with another source without rebuilding the AI, Telegram bot, and queue.

Protection against duplicates will be implemented by ID/URL of the source and an additional fingerprint from the title, date, and city. For publications, there will also be a unique key that will prevent resending the same event after an error or restart.

Regarding photos, I propose a secure hybrid option:

— automatically use a pre-prepared library of licensed sports photos or own branded templates;
— use the organizer's photo only with clear permission, preserving the source and authorship;
— allow the administrator in the Telegram bot to replace the photo by sending their own;
— not automatically copy images from Finishers or Google without checking rights.

The bot will send a complete preview with a photo, text, official link, and buttons "Approve," "Reject," and "Replace photo." After approval, the record will receive the nearest available slot. For each channel, the status will be stored: found, prepared, approved, scheduled, published, or error.

The AI will only work with pre-allocated facts. It will not invent missing data but will mark it for verification. The prompt and schedule can be changed without editing the code.

Final cost of the MVP: $600. This estimate includes one source of Finishers, one administrator, one Telegram channel, one Facebook- Expand

Budget: 20000 UAH Deadline: 12 days

Finishers does not provide a public API, so the first step is to assess the possibility of parsing the page or finding alternative channels to access their data (RSS, export, direct contact with the team).

I will create a system on n8n with built-in HTTP requests for data parsing, OpenAI for content adaptation in Ukrainian, Telegram Bot API for moderation through the bot, and Meta API for publishing on Facebook. For photos, I will offer two options: either automatic search for free images through Unsplash/Pexels API with relevant tags, or saving links to original photos with attribution in the publication's caption. The architecture will be modular — event sources will be stored in a separate Airtable table, allowing new sources to be added without changing the main workflow.

I offer 2 weeks of free support and adjustments after delivery, as the system will require testing with real data and possible corrections in the parsing logic.

I am ready to start.

Budget: 27000 UAH Deadline: 10 days

Hello, The task is completely clear. Implementation in pure Python is the best solution: it will ensure full autonomy, speed, and easy addition of new sources in the future.
Brief responses to your points:
Proposed stack: Python (OOP architecture), Playwright/BeautifulSoup (parsing), SQLite/PostgreSQL (database), OpenAI API (gpt-4o-mini for AI), Aiogram 3 (admin bot moderation), Meta API and Telegram API (auto-posting).
Approach to Finishers and alternatives: I will use a script with rotating proxies to bypass Cloudflare. It enters the event and extracts the direct link to the organizer's website. Due to the modular structure, adding or replacing a source in the future will take only 1-2 hours without touching the AI and bot logic.
Photo solution: The script automatically uploads the original promo photo from the event website (this is legal as an informational announcement with a link to the author). If the photo is missing, the AI sends a request to the free stock Unsplash API using keywords and pulls quality media. I will automatically overlay your watermark using the Pillow library.
How to make it cheaper: I will package all the code into a Docker container and deploy it on a VPS for $4-5/month. This will completely free you from expensive subscriptions to Make or n8n Cloud and limitations on the number of operations.
Estimated timeline: 10 days.
Final cost of MVP: $550.

Budget: 20000 UAH Deadline: 10 days

Hello, my stack (Python, Playwright, Selenium, aiogram, FastAPI, PostgreSQL, Docker).
I have practical experience in developing multifunctional Telegram applications with external API integration for media analysis, as well as creating robust asynchronous automation and data extraction systems.
I propose to implement an MVP on custom code instead of no-code platforms, which will ensure flexibility, reliability, and reduce monthly subscription costs.

Budget: 20000 UAH Deadline: 14 days

You need not just auto-publication, but a managed flow of events: so that the administrator quickly approves quality content, and the audience goes directly to the official registration pages, without manual searching and rewriting of descriptions.

For the MVP, I propose using n8n as the basis for the scenarios, OpenAI or Claude for cleaning and adapting texts in Ukrainian, Telegram Bot API for previews and approvals, Airtable or Google Sheets as the publication queue, and Meta API for Facebook and Telegram Bot API for the channel. I will separate the sources into a separate module, so Finishers can be replaced or supplemented in the future without restructuring the system.

I will first check the data retrieval from Finishers for validity and stability: API or available structured pages; an alternative will be Apify or connecting to the official sources of the organizers. For photos, I suggest using images from official pages only with permission or generating neutral AI visuals.

I created a system for collecting and AI processing data about potential clients using Apify + OpenAI.

The cost of the MVP is 20,000 UAH, the term is up to 2 weeks. How many events per week do you plan to publish at the start?

Budget: 20000 UAH Deadline: 21 days

Hello! I propose a solution based on n8n + OpenAI API + Telegram Bot API + Meta Graph API, with Airtable as the database for events, queues, and logs.
Experience: I have done similar automations on Make.com/n8n — specifically a lead routing system (LeadSort Pro) and a Telegram digest based on AI content processing from external sources, so the logic "source → AI processing → approval → publication" is familiar to me.
Approach to Finishers: there is no official API, so access will be through scheduled parsing of the events page with a unique identifier for each event (protection against duplicates). I allocate separate time for testing the stability of this source — this is the riskiest part of the MVP, and I honestly warn about it right away.
Photos: I propose a semi-automatic option — the system attempts to automatically retrieve images from the official organizer's website, and if not found or rights are unclear — the administrator uploads photos directly in the Telegram bot at the approval stage.
Timeline: 14–21 days (part of the time is for testing the source parsing).
MVP cost: $550–600.
Monthly support approximately $20–40 (VPS + OpenAI API based on usage, Telegram/Meta API are free).
I am ready to discuss the architecture in more detail if interested.

Budget: 27000 UAH Deadline: 10 days

Hello, Artem! I am Roman, a specialist in process automation and API integration. I have experience in creating tailored works: AI parsing of content from web sources into relevant information, setting up n8n/Make.com scenarios for large-scale mailings on Telegram and Meta platforms, working with OpenAI API to adapt content for the Ukrainian language. With past experience, I filtered thousands of events on sports topics, automated publications on social media considering image rights.

I propose a solution: first, I will set up parsing Finishers through their API or web scraping (considering limits), then AI editing of descriptions and translations through OpenAI, generating previews based on Canva API with a watermark to limit photo usage. The administrator approves the content in the Telegram bot, then automatic publication on Facebook with a link to the official website. The price is 27,000 hryvnias for the MVP, the term is 10 days. I include a free revision for 14 days after launch.

Budget: 20000 UAH Deadline: 10 days

Artem, the task is clear: you need a stable connection that pulls data from Finishers, adapts it through AI to your requirements, and automatically publishes it after moderation. The main challenge here is working with the data source and correctly processing images to avoid copyright issues. I suggest starting with an analysis of the technical feasibility of parsing Finishers to choose a method that won't be blocked, and for the content, set up automation through n8n or Make, where AI will not only translate the text but also structure it for each platform. How do you plan to address the issue of photo uniqueness for posts to minimize risks for FB pages?