Ai

Firecrawl v2.10 Adds Document Parsing and Lockdown Mode

Firecrawl v2.10 Adds Document Parsing and Lockdown Mode

Firecrawl has dropped version 2.10, and it's a big one for anyone wrangling messy documents. The web scraping tool now lets you upload local files — PDFs, Word docs, spreadsheets, you name it — and get back clean, LLM-ready Markdown. Plus, there's a new Lockdown Mode for stricter scraping controls. Here's what changed and why it matters.

What Changed

The headline feature is the /parse endpoint. It handles PDF, DOCX, DOC, ODT, RTF, XLSX, XLS, and HTML files up to 50 MB. Upload a file, and Firecrawl returns Markdown, JSON, or a summary — complete with preserved tables and reading order. For enterprise plans, there's full Zero Data Retention support, meaning your sensitive docs aren't stored. The endpoint is accessible via SDKs in JS, Python, Go, Rust, Java, .NET, PHP, Ruby, and Elixir. That's a lot of languages.

Then there's Lockdown Mode. Set lockdown: true on the /scrape endpoint, and results come exclusively from Firecrawl's index — zero outbound requests. It's a security feature for teams that need to guarantee no external network calls leak metadata. Simple toggle, big implications.

Why It Matters

Document parsing is a pain point for AI pipelines. You've got PDFs with messy layouts, spreadsheets with embedded tables, and everyone wants LLM-friendly output. Firecrawl's /parse is a direct solve. Instead of juggling separate OCR tools or custom extractors, you get one endpoint that does it all. The 50 MB limit covers most business files — think annual reports, legal contracts, or financial statements. And the Zero Data Retention option? That's a compliance must-have for regulated industries.

Lockdown Mode is more niche but equally important. For internal tools or high-security environments, you don't want the scraper wandering off to the internet. Firecrawl's index alone might be enough for many use cases. It's a simple on/off switch that changes risk posture dramatically.

The SDK expansion is smart. By supporting a dozen languages, Firecrawl lowers the barrier for adoption. Developers can integrate without fighting library quirks. That's how you build a platform play.

One caveat: no mention of pricing changes or whether /parse consumes credits differently. That's the kind of detail enterprise buyers will want. But for now, v2.10 feels like a genuine step forward — not just another patch note.

Official Source: https://github.com/firecrawl/firecrawl/releases/tag/v2.10