Step 2: Connect the Browseract Scrape Task

Configure the API call that launches the scraping task.

1. Open Launch Scrape Task and set URL to https://api.browseract.com/v2/workflow/run-task.
2. Set Method to POST and enable Send Body.
3. In Body Parameters, set workflow_id to [YOUR_ID].
4. Credential Required: Connect your httpBearerAuth credentials for Launch Scrape Task.

Step 3: Validate Task Response and Poll for Completion

Use conditional checks and wait windows to retry until the task finishes.

1. Open Validate Task Response and confirm the conditions use {{ $json.error }} with notExists and {{ $json.id }} with notEquals null.
2. From Validate Task Response, keep the true branch connected to Retrieve Task Details and the false branch connected to Retry Pause.
3. In Retry Pause, set Unit to minutes and Amount to 1.
4. Open Retrieve Task Details and set URL to https://api.browseract.com/v2/workflow/get-task with query parameter task_id set to {{ $json.id }}.
5. Credential Required: Connect your httpBearerAuth credentials for Retrieve Task Details.
6. In Verify Completion State, confirm conditions use {{ $json.error }} with notExists and {{ $json.status }} equals finished.
7. Keep the false branch from Verify Completion State routed to Pause Window, and the true branch routed to Keyword Match Agent.

Step 4: Set Up AI Keyword Matching

Configure the AI agent, language model, and structured parser to filter relevant news items.

1. Open Keyword Match Agent and keep Prompt Type as define with the provided prompt text.
2. Ensure Keyword Match Agent has Structured Result Parser connected as the output parser.
3. Open Structured Result Parser and keep the JSON Schema Example set to the provided structure for Matched_News.
4. Attach Gemini Chat Model to Keyword Match Agent as the language model.
5. Credential Required: Connect your googlePalmApi credentials for Gemini Chat Model. The Structured Result Parser uses the parent agent, so add credentials to Gemini Chat Model, not the parser.

Step 5: Format and Send to Telegram

Transform the matched items into Telegram-ready payloads and dispatch photos with captions.

1. Open Format Matched Items and keep the JavaScript code as-is to map Matched_News into { "text": item } objects.
2. Open Telegram Photo Dispatch and set Operation to sendPhoto.
3. Set File to {{ $json.text.Url }} and Caption to {{ $json.text.Headline }}

{{ $json.text.Url }}.
4. Replace Chat ID value [YOUR_ID] with your Telegram chat or channel ID.
5. Credential Required: Connect your telegramApi credentials for Telegram Photo Dispatch.

Step 6: Test and Activate Your Workflow

Validate the end-to-end flow and enable the workflow for production runs.

1. Click Execute Workflow to run from Scheduled Automation Trigger.
2. Confirm that Launch Scrape Task returns an id, then Retrieve Task Details shows status as finished.
3. Verify Keyword Match Agent outputs Matched_News and that Telegram Photo Dispatch sends photos with captions to your chat.
4. When satisfied, toggle the workflow to Active for scheduled production runs.