AI-Powered Task Decomposition from Telegram to Todoist - n8n Workflow

Individuals who manage projects on the go and prefer using voice notes for capturing ideas.
Teams looking for a seamless way to transform unstructured ideas from Telegram into actionable Todoist tasks.
Users interested in advanced n8n workflow templates leveraging OpenAI's project management capabilities.
Automation specialists seeking practical examples of complex LLM tooling within an n8n node setup.

This sophisticated n8n workflow solves the common problem of turning quick, unstructured communication (like a hurried voice note or text message captured via Telegram) into structured, actionable project tasks. By integrating a Telegram trigger with powerful AI capabilities via the OpenAI n8n node, this solution automatically fetches the message, uses the AI (specifically GPT-4o-mini) to act as a project manager, breaking down complex tasks into sub-tasks, assigning priority levels (1-4), and then feeding these structured items directly into Todoist. This is a highly effective n8n templates solution for maintaining productivity without ever leaving your chat app. Every task creation is confirmed back in Telegram, completing the loop using a Telegram n8n node.

1. Trigger Event: The Receive Telegram Messages n8n trigger initiates the n8n workflow upon receiving any new message (text or voice) in the configured Telegram chat.


2. Input Classification: The Voice or Text? switch n8n node checks if the message contains a voice file ID or plain text, routing the flow accordingly.


3. Voice Processing: If a voice message is detected, the Fetch Voice Message n8n node downloads the file. The Transcribe Voice to Text n8n node then uses the OpenAI Whisper API to accurately convert the audio into transcribed text.


4. Text Processing: If a text message is detected, the Prepare for LLM n8n node extracts the content and prepares the payload for the next step.


5. AI Decomposition: The transcribed or raw text is fed into the Basic LLM Chain. This central n8n node, leveraging the configured OpenAI Chat Model (GPT-4o-mini), follows a strict prompt instructing it to perform task decomposition. It analyzes the input and generates a list of JSON objects, where each object represents a single sub-task with content and priority.


6. Task Creation: The n8n workflow iterates over the structured outputs generated by the LLM. The Create Todoist Tasks n8n node uses the dynamically mapped content and priority fields to instantly create new, highly organized tasks within a specified Todoist project.


7. Confirmation: Finally, the Send Confirmation n8n node uses the original chat ID from the n8n trigger to send a confirmation message back to the user on Telegram, providing the task description and a link to the newly created Todoist task.

1. Import n8n Workflow: Copy the provided JSON data and import it into your self-hosted or cloud n8n instance using the 'New' -> 'Import from JSON' option.


2. Telegram Setup: Set up your Bot Token credentials for the Receive Telegram Messages and Send Confirmation n8n nodes. Ensure you activate the Telegram n8n trigger node to create the necessary webhook.


3. OpenAI Credentials: Configure your OpenAI API Key credentials for the OpenAI Chat Model and Transcribe Voice to Text n8n nodes.


4. Todoist Setup: Configure your Todoist API Token credentials for the Create Todoist Tasks n8n node. Important: In the Create Todoist Tasks n8n node settings, replace the placeholder Project ID with the actual ID of your desired Todoist project.


5. Activation: Once all credentials are linked and tested, activate the n8n workflow to begin listening for Telegram messages.

Receive Telegram Messages (n8n trigger): The primary entry point for this n8n workflow. It is configured to initiate the execution upon receiving any text or voice message in the associated chat.
Voice or Text? (Switch n8n node): A flow control mechanism that checks for the existence of message.voice.file_id (Audio path) or message.text (Text path), directing the flow to the appropriate processing steps.
Transcribe Voice to Text (OpenAI n8n node): Handles the conversion of voice messages into transcribed text using the specialized OpenAI translate operation, making unstructured audio accessible to the downstream AI n8n node.
Basic LLM Chain (LangChain n8n node): This n8n node executes the core project management logic. It utilizes a comprehensive system prompt to ensure the OpenAI Chat Model breaks down the user's request into actionable sub-tasks and strictly formats the output as a JSON list with content and priority fields.
OpenAI Chat Model (LangChain n8n node): Provides the engine for the LLM chain, configured here to use the highly efficient gpt-4o-mini model.
Create Todoist Tasks (Todoist n8n node): The final action node. It consumes the structured JSON list from the AI and creates individual, prioritized tasks in Todoist, utilizing dynamic expressions for the task content and priority from the previous n8n node's output.

• Send Confirmation (Telegram n8n node): Confirms successful execution by sending a message back to the originating Telegram chat, including details like the task content and the link ($json.url) created by the Todoist API.