Sumitr Mardy

I build Apify actors for people who call them from code, workflows and AI agents, where the caller is a machine and one unexpected null breaks everything downstream.

What every actor promises

• One flat row per input, every field always present. A failed input still comes back as a row with a machine-readable status and subStatus, so nothing disappears silently.
• Pay for results, not attempts. Pay-per-event pricing: failed, empty and skipped items are free. Prices come from costs measured on the platform and are spelled out in every README.
• Measured, not claimed. Every actor ships after a feasibility test on Apify itself. The speed, cost and accuracy numbers in the docs come from real runs, including what fails and why.
• Honest limits. What an actor cannot do is written in its README before you run it.
• Agent-ready. Every actor is a tool on the Apify MCP server and listed on the official MCP Registry, with an output schema an agent can read before the first call.

Actors

• Text from media: Audio & Video Transcriber (Whisper, podcast feeds, SRT) · Instagram Reels Transcript · Image to Text OCR · PDF Text Extractor
• Search & market data: Google Trends API · Google Ads Monitor
• Contact data: Email Verifier (SMTP, catch-all) · Website Contact Finder

New actors are added every week.

Open to custom data pipelines, scraping and verification work: reach me on GitHub.