Multi-angle podcast research across tens of millions of moments. DeepSeek V4's open-source model lets us throw massive compute at your questions without raising prices.

What you get:

• Multi-angle podcast research across tens of millions of moments

• Better proper-noun search (actually finds people and organizations now)

• Deep Mode (60-90s, thorough) or Fast Mode (30-45s, quick lookups)

• Same 10 cents per call via L402 Lightning

Why it matters:

DeepSeek V4 costs a fraction of closed models, so we reinvested those savings into deeper search. You get more comprehensive answers without paying monthly SaaS costs. Use our web app or L402 enabled API with your agent.

Open source models are changing what's economically feasible. This is what that looks like.

Try the web app: https://www.pullthatupjamie.ai/app?view=agent
Docs for Your Agent: Full docs: https://www.pullthatupjamie.ai/llms.txt
Read the full write-up: https://www.pullthatupjamie.ai/blog/jamie-deepseek-upgrade-20260428