Step 2: Route Voice vs. Text Inputs

This step branches the flow based on whether the user sent a voice note or text.

1. In Detect Voice Or Text, keep the two rules that check {{ $json.message.voice.file_id }} and {{ $json.message.text }} to create Voice and Text outputs.
2. For the voice branch, configure Retrieve Voice File with Resource set to file and File ID set to {{ $json.message.voice.file_id }}.
3. Credential Required: Connect your telegramApi credentials in Retrieve Voice File.
4. For the text branch, configure Capture Text Input to set text to {{ $json.message.text }}.
5. Connect Retrieve Voice File → Transcribe Audio and Capture Text Input → Lead Scraper Agent.

Step 3: Set Up AI Processing for Query Extraction

This step turns voice or text into a structured query that the scraper can use.

1. In Transcribe Audio, keep Resource set to audio and Operation set to transcribe.
2. Credential Required: Connect your openAiApi credentials in Transcribe Audio.
3. Configure Lead Scraper Agent with Text set to {{ $json.text }} and ensure Prompt Type is define with Has Output Parser enabled.
4. In OpenAI Chat Engine, set the Model to gpt-4.1-nano (or your preferred model).
5. Credential Required: Connect your openAiApi credentials in OpenAI Chat Engine.
6. Attach Buffer Memory and Structured Parser to Lead Scraper Agent as AI sub-nodes. Ensure credentials are added to OpenAI Chat Engine (parent), not to the sub-nodes.
7. Confirm Structured Parser uses the provided JSON schema example for location, business, and job_title.

Lead Scraper Agent outputs to both Build Query Payload and Fetch Existing Emails in parallel.

Step 4: Build and Execute the Lead Scrape

Generate a search URL, run the Apify actor, and map the lead results into a consistent structure.

1. In Build Query Payload, set Mode to raw and JSON Output to { "query": {{ $json.output }} }.
2. In Assemble Search Link, keep the provided JavaScript to build the Apollo URL and output finalURL.
3. In Execute Apify Actor, set Operation to Run actor, Build to latest, Timeout to 10000, and Wait For Finish to 60.
4. Set Custom Body in Execute Apify Actor to { "getPersonalEmails": true, "getWorkEmails": true, "totalRecords": 500, "url": "{{ $json.finalURL }}" }.
5. Credential Required: Connect your apifyApi credentials in Execute Apify Actor.
6. In Map Lead Details, map fields like firstName to {{ $json.first_name }}, emailAddress to {{ $json.email }}, and jobTitle to {{ $json.employment_history[0].title }}.

⚠️ Common Pitfall: If the scraper returns empty employment_history, the expression {{ $json.employment_history[0].title }} will fail. Consider adding guard logic if needed.

Step 5: Filter and Deduplicate Leads

Only verified emails are saved, and existing emails are removed before insertion.

1. In Filter Verified Emails, keep the condition emailStatus equals verified using {{ $json.emailStatus }}.
2. In Fetch Existing Emails, set Operation to select, Return All to true, and include outputColumns of emailAddress.
3. Credential Required: Connect your postgres credentials in Fetch Existing Emails.
4. In Isolate New Leads, keep Merge By Fields configured to compare incoming emails with existing ones.
5. Leave Marked Existing as a no-op branch for duplicates.

Step 6: Store Results and Send Telegram Confirmation

New leads are inserted into Supabase, a confirmation message is composed, and a Telegram reply is sent.

1. In Insert New Leads, set Table ID to Leads_n-mail and map fields like firstName to {{ $json.firstName }}, jobTitle to {{ $json.jobTitle }}, and country to {{ $json.country }}.
2. Review the businessIndustry mapping: {{ $json.businessIndustry }}{{ $('Apollo Scraper').item.json.organization.industry }}.
3. Credential Required: Connect your supabase credentials in Insert New Leads (no credentials are configured yet).
4. In Compose Telegram Reply, set output to {{ $input.all().length }} new contacts have been added to the Google Sheet!.
5. In Send Telegram Confirmation, set Text to {{ $input.all().length }} new contacts have been added to the Google Sheet! and Chat ID to [YOUR_ID].
6. Credential Required: Connect your telegramApi credentials in Send Telegram Confirmation.

⚠️ Common Pitfall: The workflow references Apollo Scraper in Insert New Leads, but that node does not exist in this workflow. Replace that reference or remove it to prevent execution errors.

Step 7: Review Placeholder and Non-Functional Nodes

This workflow includes non-executing nodes that are used for annotation or future expansion.

1. Keep Flowpast Branding as a documentation note in the canvas.
2. Identify Unnamed (type unknown) and remove or replace it if it appears in any future connections.

Step 8: Test and Activate Your Workflow

Validate the workflow end-to-end and then enable it for production use.

1. Click Execute Workflow and send a text or voice message to your Telegram bot.
2. Confirm that Detect Voice Or Text routes correctly and that Lead Scraper Agent outputs a structured query.
3. Verify Execute Apify Actor returns lead data and that Insert New Leads creates records in Leads_n-mail.
4. Check that Send Telegram Confirmation delivers a message with the correct count.
5. When successful, toggle the workflow to Active to run automatically on new Telegram messages.