Parsing and then inviting users from Telegram chats

Technical Task: Context The foreign language studio "AMI" (Vinnytsia, 10 years in the market) is launching a short pre-recorded online foreign language course. The course is accompanied by an AI assistant named Starky — the studio's symbol. Starky acts as a "digital tutor" that supports the student between course lessons. The course is hosted on the WeStudy platform (Elite tariff).Project Goal To implement a chatbot based on a language model (LLM) that: responds to the student on behalf of the character Starky (a friendly, encouraging cat assistant); helps with questions related to the current lesson topic; supports the student's motivation (reminders, praise for progress); does not answer questions outside the course (general topics, other subjects, etc.).Functional Requirements Interaction Channel: Telegram bot (preferred option — quickly implemented, familiar interface for students). If there is a technical possibility to embed a chat widget directly on the WeStudy platform — we will consider it as an alternative, but Telegram is the basic scenario. Connection to LLM: via API (for example, Anthropic Claude API or similar at the contractor's choice, with justification for the choice). System Prompt / "Character" of Starky: configuration of instructions that define the tone of communication, allowed and prohibited topics. The client provides a description of the character's personality and behavior rules. Link to Course Content: the bot must have access to materials from 5–10 course modules (lesson texts/notes) to respond specifically to the current module topic, rather than general phrases. Simple Administration: the ability to update course materials and the system prompt without involving a developer (for example, through a text file or a simple admin panel). Testing: verification of the correctness of responses to 15–20 typical student questions before launch.What is NOT included in this Technical Task Development of the course content itself (lessons, exercises) — to be done by the studio independently. Design/branding of the Starky character (visual image) — to be discussed separately if needed. Ongoing long-term support after launch (separate agreement if needed).Expected Results (deliverables) Working Telegram bot connected to LLM API. Configured system prompt for the Starky character. Instructions for the client: how to update course content and adjust the prompt. Brief report on testing.Estimated Budget and Timeline Budget: up to 25,000 UAH (maximum amount agreed within grant funding; please indicate your estimate — fixed price or hourly rate with an estimate of hours). Implementation Timeline: preferably within 4–6 weeks from the moment of confirmation, agreement — separately. Additionally, please indicate the estimated monthly cost of using the API (payment for tokens/messages) — this is a separate expense item from one-time development.

I need to create an automated SEO system based on n8n that will collect data from various services, combine it, and send it to ChatGPT for analysis. It is necessary to connect Ahrefs Google Search Console Google Analytics 4 Google Ads The first stage is a minimum viable version Connect Google Search Console. Connect GA4. Connect Ahrefs, if my plan and API allow it. Store daily data. Send it to ChatGPT. Send one daily report to Telegram or email. Set up analysis for the site viarcanvas.com and the competitor picturehappy.lv. After verifying the functionality, add Google Ads, additional countries, competitors, and extended reports. What needs to be obtained from Ahrefs For the site viarcanvas.com and selected competitors: organic keywords; current positions; position changes; search volume of queries; estimated traffic; pages that receive organic traffic; new and lost keywords; new and lost backlinks; competitor pages that have increased traffic; new competitor keywords; Content Gap between our site and competitors. The list of competitors should be editable without changing the n8n script itself.What needs to be obtained from Google Search Console clicks; impressions; CTR; average position; queries; pages; countries; devices; changes in metrics by days and periods. It is important that the data can be analyzed separately for each language version: /lv/ /lt/ /ee/ /pl/ /de/ A separate analysis of specific pages is also needed, for example: https://viarcanvas.com/lv/new/canvasWhat needs to be obtained from Google Analytics 4 users; sessions; organic traffic; landing pages; purchases; leads; revenue; conversion rate; engagement; countries and devices. It is necessary to link SEO traffic not only to visits but also to leads, orders, and revenue.What needs to be obtained from Google Ads spending; clicks; impressions; CTR; CPC; conversions; cost per conversion; revenue; ROAS; search queries; data on campaigns and countries. Google Ads should be analyzed separately from organic SEO but included in the overall marketing report.What ChatGPT should analyze ChatGPT should automatically determine: which keywords have increased; which keywords have decreased; which pages have lost traffic; which pages have started to grow; where impressions have decreased; where CTR has decreased; where the average position has worsened; which pages are in positions 4–15 and have the potential to reach the TOP; which keywords have impressions but few clicks; which competitor pages are growing faster than ours; which queries competitors have that we do not; which articles should be written; which existing pages should be updated; where query cannibalization is possible; where internal linking is needed; which pages should be strengthened with external links; which changes require urgent attention. ChatGPT should not just list numbers. It should explain the likely reasons for changes and provide a specific action plan.

https://docs.google.com/document/d/1TpVG_3eorqptb4MCf5XmcvwXJMuAA56UFyusw_7wAfI/edit?tab=t.q8f7pv3dj31d

Technical specification for a Telegram bot Goal: Create a Telegram bot that will: 1. Take posts from one channel (source). 2. Repost them to one or several other channels (receivers) on behalf of the Telegram account. 3. Support: • Text. • Up to 10 photos. • Premium emojis. 4. Replace links in posts according to specified rules. 5. Add inline buttons according to special scripts in the text. 6. Support link and button binding for each receiver separately. Working with links: • There may be links in the post text that need to be replaced. • For each receiver channel, the bot should use its own link. • Example: if the post contains the link https://site.com, then: • For Channel A replace with https://site.com/a • For Channel B replace with https://site.com/b Working with buttons: • The text may contain scripts like: [contact], [website], [order]. • The bot should: • Remove the script from the text. • Add a button with the required text and the required link. • Buttons are configured separately for each receiver channel.

Development of a Telegram bot in Python that collects data and synchronizes it with Google Sheets in real time. The system must be fault-tolerant, protected against manual input errors, and fully autonomous.