Why we built Sessemi

April 2026 · Andrew Odiit

Sessemi exists because of a pharmacy chain in France.

I was scraping Aprium — over a million products a week, hundreds of storefronts. It was a production pipeline: reliable, scheduled, running for months. Then one week the data stopped coming. Aprium had deployed Cloudflare Turnstile with geo-blocking. Every request hit a challenge page. The pipeline went from a million products to zero overnight.

This wasn't a new problem. I'd been scraping professionally for years, building infrastructure that fed real business decisions. Anti-bot protection was always part of the job. But this time, the tools I'd been relying on couldn't keep up.

What I tried first

ScrapingBee doesn't solve challenges. It proxies the request and returns whatever the server sends back — which, when Cloudflare is in the picture, is the challenge page. Useful for unprotected sites, useless for this.

Capsolver is a cookie farming service — it solves challenges in its own environment, then hands you back the clearance cookie to replay in your scraping client. It worked against basic Cloudflare setups. But Aprium had Turnstile with geo-blocking and aggressive TLS fingerprinting. When Cloudflare enforces TLS fingerprinting, cookies are bound to the connection that created them. A cookie solved in Capsolver's environment and replayed from a different TLS stack gets rejected. It broke at scale.

Scrapfly I didn't bother trying. At a million products a week, the per-request pricing made it dead on arrival. I did the math, closed the tab.

The open source path

I looked at everything: headless Chrome, Camoufox, undetected-chromedriver, Playwright with stealth plugins. The tooling has gotten better, but the fundamental problem hasn't changed — you're beholden to the publishers of these systems.

When Chrome ships a new version, you wait for the stealth patches. When Cloudflare changes their detection, you wait for the community to figure it out. When a project's maintainer moves on, your production pipeline is running on abandoned code. The lag between "anti-bot vendor ships an update" and "open source tool catches up" can be days or weeks. In production, that's downtime.

It's the same dependency problem as the API vendors, just with a different tradeoff. You save money but lose reliability. For a pipeline doing a million requests a week, neither is acceptable.

Starting from scratch

As an engineer, my job is to solve problems. When the existing solutions don't work, you build a new one.

I decided to build the entire anti-bot bypass and solving system from scratch — detection, challenge solving, session management, proxy orchestration — written entirely in Go, with no third-party CAPTCHA services in the loop. No dependency on Chrome's release cycle. No waiting for someone else to push a patch when a vendor changes their challenge format.

The result is Sessemi. One API call. If the site is protected by Cloudflare, DataDome, or Akamai, Sessemi detects it and solves it — automatically, on every request.

curl "https://api.sessemi.com/scrape?url=https://protected-site.com&solve=true" \
  -H "Authorization: Bearer YOUR_KEY"

First request to a domain pays a solve cost — a few seconds while the challenge is handled. Subsequent requests are fast, typically under two seconds. For batch workloads, that's one solve and 99 fast responses.

No per-solve fees

Every competing API I evaluated charges extra for challenge solves — per-solve pricing, or a 10x credit multiplier for "protected pages." Sessemi has flat credit costs. A protected page costs more credits than an unprotected one, but the solve is included. Your bill doesn't spike because a site added Cloudflare overnight.

Why Go?

There's a reason Sessemi is written in Go and not Python or Node. It's a deliberate architectural choice that affects performance, reliability, and what the system can do under load. That's a post on its own — coming soon.

Try it

Sessemi is live. The free tier gives you 500 credits — enough to test against your actual targets before committing. The changelog shows what we're shipping and when.

If you're scraping sites behind Cloudflare, DataDome, or Akamai, and you're tired of paying for solved CAPTCHAs that don't work, give it a try.