Development of an AI bot for receiving and processing orders from restaurants in Telegram with integration into 1C.

Budget: 21000 UAH Deadline: 5 days

Hello. I am implementing a bot in Python (Aiogram + FastAPI) with a full order processing cycle.

For recognition, I will use Whisper (voice) and OpenAI Vision / EasyOCR (photo of handwriting). I will build the logic for slang mapping (“картоха” -> “Картопля”) and entity extraction through the OpenAI API (Structured Outputs/JSON mode) with strict validation of pydantic schemas. For integration with 1C, I will set up REST API integration (1C HTTP services) with status and error handling.

Budget: 20000 UAH Deadline: 4 days

Good day. I am ready to complete your order. I have experience in developing bots using Aiogram + Whisper.

Budget: 20000 UAH Deadline: 5 days

Hello!

I develop bots for Telegram using NodeJS. I'm ready to take on the project. Write to me, and we will discuss.

Budget: 20000 UAH Deadline: 7 days

Hello, I recently created a similar Telegram bot for order processing: handling free text and voice through Whisper, recognizing items and quantities, matching customer slang with the official product catalog. Clarifying questions in the chat and final confirmation before sending the order were also part of that project.

If you have many regular restaurants, it would be worth adding memory for matches for each customer, so the bot guesses "potato" the first time and doesn't ask again, which eliminates up to 30% of clarifications.

How many items are in your catalog and how is 1C currently available: are there ready REST services or do they need to be set up? And where does the bot get the product catalog from, is it an export or a direct query?

I suggest we connect, and I will also outline the processing scheme for the request from voice to the document "Customer Order" in 1C for your catalog.

Budget: 19000 UAH Deadline: 5 days

Hello! I will do it quickly and efficiently. I am ready to tackle your task. Write to me, and we will agree.

Budget: 20000 UAH Deadline: 10 days

Good day.
Our team has many years of experience in developing ERP, CRM, CMS, and specialized software for businesses. We create effective digital solutions that help automate processes, increase productivity, and scale companies.

We work with modern technologies — from bots and scripts to AI agents and analytical systems. We develop websites of varying complexity. Our portfolio includes implemented ERP solutions for the hotel business, as well as for companies engaged in the import and sale of goods, and our own product XFitness — an ERP system created specifically for fitness clubs.

We are ready to implement your project and offer the best solution tailored to your needs.
Our portfolio: Freelancehunt

We specialize in the following areas:
- Development of ERP Systems
- Development of CRM Systems
- Development of Websites of any complexity
- Development of CMS Systems
- Support for Websites
- Development of OpenCart
- Support for OpenCart
- Modification of OpenCart
- Enhancement of OpenCart
- Development of WordPress
- Support for WordPress
- Modification of WordPress
- Enhancement of WordPress
- Development of ECommerce
- Support for ECommerce
- Modification of ECommerce
- Enhancement of ECommerce
- Development of Web Applications
- Support for 1C Servers
- Support for Web Servers
- Development of mobile applications
- Data parsing
- Development of bots
- Development of AI agents

and on the following technologies:
- Python
- PHP
- Laravel
- Symfony
- Yii2
- JS
- NodeJS
- jQuery
- TypeScript
- MySQL
- HTML
- CSS
- Vue
- Nuxt.js
- React
- React Native
- C++ Expand

Budget: 20000 UAH Deadline: 14 days

Good day, I am a Python developer. I have extensive experience working with Telegram bots, integrating with various APIs, and also with 1C. I developed a Telegram bot that works with 1C in the form of MCP with Claude to retrieve and add various data; I can show examples of such bots. I am ready to complete your order quickly and efficiently. Write to me - we will discuss the details, I am ready today.

Budget: 19000 UAH Deadline: 25 days

This is an interesting project, and the most challenging part here is definitely not whisper or OCR, but the mapping of nomenclature. "Potato," "onion," "eggplants," this is an endless living slang, and if the mapping is done as a static table, it will become outdated in a month. So the first question is: do you already have any nomenclature reference from 1C that can be exported? The structure of it greatly affects the approach to embedding similarity search, rather than just fuzzy matching.

Regarding 1C, everything is indeed possible through REST/HTTP services, but it's important to understand which version of the configuration is being used and whether the API is open externally. There have been situations where there is a custom configuration and standard endpoints do not fit.

Overall, I see this as a job for about 25 days, taking into account all three stages and proper coverage of edge cases. Please let me know if there is already an active bot or if we are starting from scratch, and what version of 1C you are working with?

Budget: 20000 UAH Deadline: 16 days

Good day. I understand the technical specifications: a Telegram bot instead of an operator — it accepts text/photo/voice, maps to your nomenclature, and creates a "Customer Order" in 1C. I will address key risks:
- unstructured input (slang, handwriting, voice) → structured JSON (item, quantity, unit of measure);
- questionable lines — clarifying questions in chat, without "silent" errors in 1C;
- confirmation from the client before sending;
- reprocessing on API failure without restarting the entire scenario.
I will do it in Python (Aiogram + FastAPI): stage 1 — parsing to JSON; stage 2 — dictionary/mapping of nomenclature; stage 3 — HTTP/REST in 1C.

Experience: I have already created Telegram order bots in retail with export to 1C (XML/API), admin panel for nomenclature/points; separately — production AI (OpenAI + function calling) and OCR/vision for documents into structured fields. I know the operations from the inside (receipts, pricing, reports) — so I build for the real flow of applications.
Clarifications before evaluation:
1) How many items are in the directory and is there already a "slang → SKU" at least partially?
2) 1C: what configuration is it and is there already an HTTP service for "Customer Orders," or does that need to be described as well?
3) Language of applications: only Ukrainian or also Russian/mixed?
4) MVP: only text at first, or immediately photo + voice?
After the answers, I will provide the timeline and price by stages. I am ready to start after agreement.

I cannot provide a link to an identical bot as it is confidential information.
https://gooru-lab.github.io/portfolio/ Expand

Budget: 20000 UAH Deadline: 14 days

Hello, Pavlo.

In this project, the critical point is not the OCR or Whisper itself, but the protection against erroneous orders in 1C. Therefore, I will build the processing as a controlled pipeline: text, voice, or photo → structured items → matching with the nomenclature → clarification of ambiguities → confirmation by the client → transfer to 1C.

Implementation: Python, aiogram, FastAPI, OpenAI Vision/Whisper, Pydantic validation, and hybrid mapping through a dictionary, fuzzy matching, and embeddings. For each uncertain item, the bot will not "guess," but will ask for confirmation. I will add a processing log, retries for integration, and idempotency to ensure that one order is not created in 1C twice.

For a correct start, I need a sample nomenclature, 15–20 anonymized real requests, and information about 1C: version and whether there is already an HTTP/REST service for creating a "Customer Order." If the endpoint is ready, I will connect it to the main cycle; if it still needs to be developed within the 1C configuration, we will fix this block separately after reviewing the structure.

Budget: 27000 UAH Deadline: 10 days

Good day.
I am a senior engineer with over 10 years of experience.
The task is quite clear.

I have experience in text recognition, transcription, and vision - I fully understand which model to choose and why (both from free open-source models and paid ones via API like Claude Opus or gpt).
I have been working with Telegram bots for quite a while using Python/NodeJS/Go.
Regarding 1C - I have experience at the level of integration in a hypermarket.

I charge a bit more than what is offered here, but by working with me, you can be assured of meeting deadlines and the quality of the final result. After completion, I will provide you with all the necessary documentation and hand over all the work as required. Additionally, I will provide technical support for a while after the launch.

Budget: 20000 UAH Deadline: 21 days

Good day!

For this task, it is advisable to build not "one AI bot," but a connection of three parts: a Telegram bot for dialogue, a backend on FastAPI for processing and storing states, and a separate integration module for 1C. Then, at the MVP stage, it will be possible to reliably accept text, voice, and photos, convert them into a unified JSON representation, and only then initiate mapping to your nomenclature and sending to 1C.

From a technical point of view, it is best to immediately establish an intermediate normalization layer: speech/image recognition should not directly create a document in 1C. After extracting positions, quantities, and units of measurement, the system should check for ambiguous fragments, ask clarifying questions in the chat, and only after confirmation, generate the order. This significantly reduces errors, especially with slang, abbreviations, and handwritten lists.

There are two critical risks. The first is the quality of OCR/ASR on real photos and voice from clients: without examples, it is difficult to guarantee accuracy, especially for handwritten text. The second is integration with 1C: it depends on whether there is a ready HTTP/REST service, what fields are needed for the "Customer Order" document, and how duplicates and synchronization errors are handled. For such integrations, it is important to include logging of each step and idempotency to avoid creating duplicate orders.

Within your budget of 20,000 UAH, it is realistic to launch the MVP stage: the bot, basic AI processing, JSON structure, clarification of unclear positions, and preparation for transfer to 1C. If the 1C API is already ready and there are examples of real orders, then the scope and timelines can be more accurately defined; if not, it is better to divide the scope into stages to avoid risking quality.

We need three things for an accurate assessment: whether you have examples of real requests in text, voice, and photos; whether there is a ready API/HTTP service in 1C with a description of document creation; and whether there is an up-to-date nomenclature reference with synonyms, abbreviations, and name variants from clients. Expand

Budget: 20000 UAH Deadline: 12 days

Hello. The described scope should be divided into separate stages with their own acceptance criteria.

For a budget of 20,000 UAH, I propose to implement the first MVP: the Telegram bot accepts text, voice, and photos, recognizes positions, quantities, and units of measurement, asks clarifying questions, and forms a confirmed structured JSON. The result is verified against an agreed set of real anonymized applications.

The comparison with the full nomenclature and integration with 1C should be evaluated separately after receiving the version and configuration of 1C, the documentation of the HTTP service, and an example of the product directory. If 20,000 UAH is the budget for the entire turnkey solution, it is impossible to implement the declared scope qualitatively within that limit.

I have 19 years of experience in industrial automation, system integration, and working with Python/API. I am ready to start with a brief agreement on the JSON schema and acceptance criteria for the first stage.

Budget: 19000 UAH Deadline: 5 days

Good day, Pavel!
On Freelancehunt, my profile is new, so I am ready to complete the first project at a reduced cost. At the same time, I have practical experience integrating 1C with my own applications, Telegram bots, and external services via API.
I can start your task immediately. Within the first 24 hours, I will show the working chain:
message, voice or photo → recognition → structured order → matching with the nomenclature → confirmation by the client.
The bot will not independently guess a questionable position. In case of ambiguity, it will ask a clarifying question, and before sending to 1C, it will show the final order to the client. This will reduce the risk of incorrect goods, quantity, or unit of measurement.
In the following stages, I will connect:
— recognition of voice messages and images;
— matching slang and abbreviations with the nomenclature;
— saving confirmed matches for regular clients;
— creating a customer order in 1C;
— logging and error handling;
— testing the full scenario with your examples.
The cost is 16,000 UAH. The deadline is up to 5 days. The first working prototype will be ready in 24 hours.
The deadline is valid if an HTTP service or API for creating an order is already prepared in 1C. If not, within the same 24 hours, I will show the completed part of recognition and order formation, and after reviewing the database, I will determine the scope of adjustments needed for the exchange with 1C.
Before starting, I just need to clarify one thing: is there already an API/HTTP service in 1C for creating the document "Customer Order," or does it also need to be developed? Expand

Budget: 27000 UAH Deadline: 10 days

Good day!

I have reviewed the task. The key objective here is not only to recognize text, voice, or photos but to produce a correct order that can be transferred to 1C without manual intervention.

I suggest implementing the project in stages.

In the first stage, the bot will accept text messages, voice messages, and photos, convert them into a structured order, and, if necessary, clarify any unclear items or quantities.

Next, a module will be added to match names with your product directory, after which integration with 1C will be connected via API.

This approach allows testing each stage separately and avoids the situation where the entire system depends on one large release.

I would like to clarify a few points:

Is there a ready REST API or HTTP services for 1C?
How many items does the nomenclature directory contain?
Are there examples of real restaurant requests (text, voice, photo) to immediately check the quality of recognition? Expand

Budget: 20000 UAH Deadline: 14 days

Bot for automating requests from restaurants: processing text, voice, photos, matching with the nomenclature, sending to 1C. I will do it in Python + Aiogram: Whisper/SpeechKit for voice, Vision OCR for photos, FastAPI for integration with 1C via REST. I will add a clarification mechanism if the data is ambiguous.

What integration format with 1C do you use: REST API or HTTP services?

Budget: 20000 UAH Deadline: 10 days

Hello. I develop Telegram bots in Python (Aiogram, FastAPI) with integration of OpenAI API (GPT-4o / Claude), Whisper for voice, and Vision models for image recognition. I have experience implementing algorithms for matching unstructured text with nomenclature directories and integrating bots with external accounting systems via REST API.

For a quality launch within a budget of 20,000 UAH, I propose to implement the project in stages:

- MVP (Stage 1): logic of the Telegram bot, processing of text and voice (Whisper), recognition of quality photo lists (Vision), extraction of entities (product/quantity) and an algorithm for matching with slang (Fuzzy matching / Embeddings) with confirmation output to the client and formation of structured JSON.
- Stage 2: Connecting data transfer to 1C and fine-tuning recognition of complex handwritten text (after testing the first stage).

Questions regarding the project:

- Do you have a 1C specialist who will provide a ready REST API / HTTP service for creating the "Customer Order" document?
- How many active positions are in your nomenclature directory?
- Are the API keys for recognition services (OpenAI / Yandex / Anthropic) provided by you?

I am ready to discuss the details of implementation and clarify technical points in personal messages, feel free to write. Expand

Budget: 19000 UAH Deadline: 7 days

Hello, Pavlo! I studied your technical specification with great interest. I am specifically offering a minimum price (undercutting) since I recently registered on Freelancehunt, and it is critically important for me to receive my first positive review in my profile and build my portfolio. I guarantee the quality of my work at 100%.

I develop asynchronous bots in Python 3.12 (aiogram 3.x), constantly integrating OpenAI/Claude API and I am skilled in working with REST requests.

I will implement the system based on reliable engineering principles, eliminating technical risks:
1. Multimodal pipeline (Whisper + Vision OCR): Voice messages are converted to text via Whisper. Photos of handwritten lists are processed through Vision OCR. To avoid errors due to poor handwriting from restaurateurs, I will set up strict prompt engineering with LLM to extract strict entities (position, quantity, unit of measure) into structured JSON.
2. Hybrid mapping of nomenclature: Slang ("kartokha", "cheri") will be matched with your official product directory through an SQLite database. Additionally, we will establish a memory of matches for each restaurant (if a client confirms once that "kartokha" means "Grade A Potatoes," the bot will remember this and will not reprocess it).
3. Verification screen (Protection against 1C errors): Anyone who promises 100% accuracy of OCR for handwritten text is lying. The working solution: the bot first outputs the recognized order to the client with inline buttons ("Is everything correct?"). The error is caught by the client in the chat, not at the warehouse.
4. Integration with 1C: After confirmation, the bot instantly sends clean JSON via HTTP service to create the "Customer Order" document.

I will deploy the bot for free on the Railway server through a secure .env configuration of keys. I am ready to start with Stage 1 (prototype for recognizing your real photos/voice into JSON) in 2-3 days.

Please let me know if your 1C already has HTTP services ready for receiving orders, or if they need to be developed together with your 1C specialist? Expand

Budget: 27000 UAH Deadline: 9 days

Good day
I have extensive experience with all the technologies required to complete this task
Write to me

Budget: 20000 UAH Deadline: 4 days

Good day! I am ready to develop such a system based on Google Sheets! I have experience in creating similar systems! It's in the reviews!!

Budget: 20001 UAH Deadline: 13 days

interesting task, ready to complete it. in a couple of weeks there will be a fully working bot. I have worked with OpenAI API and cloud, voice is not a problem. there may be nuances only with handwritten text.

Budget: 20000 UAH Deadline: 14 days

Good day! The SDEV team has extensive experience in creating complex Telegram bots and integrating them with business systems, including 1C. We implement the processing of unstructured data through AI, which will allow for the automation of order acceptance and eliminate the human factor. We guarantee the stability and speed of your solution. We are ready to discuss the details and start working soon.

Budget: 20000 UAH Deadline: 10 days

Hello. I fully understand the task of creating a Telegram bot for automating the acceptance and processing of unstructured requests for the supply of vegetables and fruits, with subsequent integration into 1C. To implement this project, I will develop a modular architecture in Python that combines advanced AI modules for multimodal recognition (text, voice, images) using LLM and OCR, with an effective nomenclature normalization mechanism. This will ensure accurate data matching and feedback to the client in case of ambiguities. With my experience in similar projects, including deep integration with corporate systems, I can quickly deploy a reliable API gateway for secure transmission of structured orders directly into 1C. I suggest discussing all implementation details, final budget, and timelines in private messages.

Budget: 20000 UAH Deadline: 3 days

Interesting project, I love tasks involving recognition and matching of nomenclature the most.

I see the task like this: I receive text, voice through Whisper, and photos of handwritten lists through Vision OCR, extract positions, quantities, and units of measurement in JSON, then match slang names like "potato" or "turnip onion" with the reference guide, and if something is unclear, the bot asks in the chat before passing it to 1C.

I am ready to start with Stage 1 and show a recognition prototype in 3 days, then mapping and integration with 1C.

Budget: 20000 UAH Deadline: 13 days

Good day, Pavlo.

First and foremost, the main question: how exactly is your 1C accessible — HTTP service, OData, or file exchange? This will determine whether the integration will take three days or if it will be a separate task along with your 1C specialist. It is impossible to provide an honest timeline without this answer.

Secondly, recognition involves three different tasks. Text and voice (Whisper) provide high accuracy. A photo of a handwritten list does not, and anyone who promises you 95% here is misleading you. The working solution: the bot shows the recognized text and asks for confirmation with one click, and only after that does the order go to 1C. The error is caught before it reaches the warehouse, not after.

Thirdly — matching customer slang with your nomenclature. This is my main strength: I wrote the hybrid search (keywords plus semantics) by hand, without ready-made libraries.

I propose to divide it: the first stage — bot, recognition, matching, confirmation, resulting in a structured order. I will demonstrate this on your actual requests. We will assess the integration into 1C after your response to the first question.

github.com/kamedashe Expand

Budget: 20000 UAH Deadline: 5 days

The main difficulty here is not in the bot on Aiogram itself, but in consolidating three input sources into a single logic: voice through Whisper, photos of handwritten lists through Vision OCR, and free text, followed by parsing entities into structured JSON.

I would do it this way: a separate recognition module, then a layer for matching slang ("kartokha," "cheri") with a nomenclature reference through embeddings plus a synonym dictionary, and only then confirmation and sending to 1C via HTTP services. For the MVP in the first stage, I will implement recognition and output JSON to immediately check the quality of position and quantity extraction.

Approximately 5 days for the first stage. Question: Does 1C already have ready HTTP services for creating an order document, or will they still need to be implemented from the 1C side?

Budget: 20000 UAH Deadline: 10 days

Hello 👋
I have reviewed the task 👀 I am currently working on a Telegram bot for 180+ employees.
Write to me, and we will discuss the details of your project! I will create a clear action plan in 1-2 hours ⏳!

Examples of websites:
https://sushi-drive.mk.ua
https://ealcoach.com
https://ms.mk.ua

----------------------------------
Website development in Wordpress / OpenCart and AI automation. Expand

Budget: 20000 UAH Deadline: 30 days

Hello. I propose to implement the system in stages: first, a Telegram bot for receiving text requests, matching with the nomenclature, clarifying undefined positions, and transferring the confirmed order to 1C. After stabilizing the main scenario, we can connect voice messages and photo recognition. For an accurate estimate, we need to review the nomenclature format and the 1C API documentation. The estimated timeline for the MVP is 20–25 working days. The budget is 20,000 UAH for the agreed scope of the first version.

Budget: 20000 UAH Deadline: 10 days

Hello!
I have experience in creating Telegram bots, integrating APIs, working with databases (SQLite/PostgreSQL), AI services (Whisper, OCR), as well as automating business processes.

Your project is completely clear. I can implement:

- receiving text, voice, and photo requests;
- recognition through Whisper and OCR;
- mapping products to your nomenclature;
- clarifying questions in case of ambiguities;
- order formation;
- integration with 1C via API;
- logging and stable bot operation.

I write clean code, maintain communication during development, and regularly demonstrate progress.

I would be happy to discuss the project details and offer an optimal solution. (I can send a personal message regarding the development of the latest Telegram bot). Since I am just starting, the price will be slightly lower than stated. Expand

Budget: 20000 UAH Deadline: 10 days

Hello! Working with photo list recognition often results in errors due to handwriting, so it's critical to build flexible prompt engineering for LLM to avoid chaos in 1C. Previously, I developed similar solutions for parsing orders and reconciling client slang with nomenclature.

Do you have a ready API from 1C for testing integration?

I suggest we discuss the architecture of the MVP.

Budget: 19200 UAH Deadline: 14 days

Pavlo, the task is clear: the bot should accept requests in any form (text, voice, photos of handwritten lists), recognize items and quantities, map them to your nomenclature, and after confirmation, create an order document in 1C via REST.

I usually work with Node.js/TypeScript (NestJS), not Python, so I propose this stack: aiogram will replace telegraf/grammy, text recognition and structuring will be done through the OpenAI API, voice through Whisper, and photos of handwritten lists through Vision. I will create a separate module for mapping client slang to the product directory with logic for clarifying questions when the item or quantity is ambiguous. The final step: a summary to the client for confirmation and sending structured JSON to 1C via an HTTP service.

It is realistic to complete all three stages described in 14 days: bot logic and recognition, nomenclature matching, integration with 1C.

A similar case: I created a Telegram bot for booking a barbershop with integration into the CRM Altegio, where there was also logic for recognizing client requests and transmitting structured data to an external system via API.

Please let me know if the 1C REST API is already set up and if there is documentation for creating an order document, or if that is also part of the work?

Budget: 25000 UAH Deadline: 20 days

Hello! I am ready to implement a Telegram bot for order acceptance with text, photo, and voice recognition through neural networks and integration with 1C.

I have experience in developing bots on Aiogram with processing unstructured data and synchronization with external systems via HTTP API.

Is there test access to the 1C API to clarify the format for creating a Customer Order document?

Example case using Google API: https://freelancehunt.com/showcase/work/telegram-bot-integrovanim-gemini/2007676.html

Budget: 20000 UAH Deadline: 4 days

I will outline the application processing as a pipeline: Whisper for voice, Vision OCR for photos of handwritten lists, parsing text into structured items (product, quantity, unit).

I will match using a nomenclature reference with fuzzy search for slang like "potato," "cherry." If the item is ambiguous, the bot will ask the client again, and after confirmation, I will send the order to 1C via REST API as a document "Customer Order."

Stage 1 (text/voice/photo in JSON) is expected to be ready in about 4 days.

Budget: 20000 UAH Deadline: 20 days

Pavlo, I understand the task: you need to minimize errors in manual order processing by automating the conversion of "live" restaurant requests into structured data for 1C. The main challenge here is working with fuzzy nomenclature and recognizing different types of input data.

To implement this, I will set up a pipeline that processes voice messages and photos through a recognition API, followed by mapping through fuzzy search using your product directory. Each recognition stage will have a verification block where the bot will ask the client again if the confidence in recognizing the product is below a specified threshold.

Do you already have a ready API interface or exchange module for communication with 1C, or should we consider developing an intermediate API on your side?