All articles
ResearchMarch 12, 2026· 8 min read
An autopsy of a Claude Code deep-research run
We traced a long deep-research session step by step to see where the time and tokens went.
PN
Priya N.
Developer Relations
We pointed a deep-research agent at a hard question, gave it a Ferr browser, and recorded everything. The trace is a great lens on where a browser agent spends its effort.
Where the time went
- Most steps were navigation, not reasoning
- Retries clustered on a handful of slow pages
- Extraction was cheap once the page was loaded
The lesson
Speeding up a research agent is mostly about speeding up the browser: faster starts, fewer retries, cleaner pages. That is squarely where Ferr helps.
Build it on Ferr
Launch your first cloud browser for free.
Keep reading
All articlesResearchMarch 19, 2026
We rebuilt the browser-agent leaderboards
A fairer, reproducible way to compare browser agents on real tasks.
Read article7 min read
ResearchSeptember 18, 2025
Benchmarking remote browsers
How we measure session start time, throughput, and stability — and what the numbers say.
Read article7 min read
Launch WeekJune 27, 2026
Launch Week v3: everything we shipped
Five days, five launches. Here is the full recap of Launch Week v3, from faster cold starts to Ferr Skills.
Read article6 min read