TL;DR: Jamie is live on PPQ. Ask "what did Saylor say about inflation last quarter" right inside PPQ — Jamie turns that into a research plan, searches millions of podcast transcript paragraphs by meaning (not keywords), and hands back the actual spoken clips with timestamps. Same Lightning balance you already use for GPT or Claude. Roughly $0.10 a call, 5–30 seconds, zero setup, no hallucinated quotes.
PPQ: Privacy and Optionality Without the Receipt
Most AI front-ends want you logged in, on a card, on a plan. PPQ doesn't. You top up with Bitcoin, Lightning, crypto, or a card if you prefer — and from there it's pay-per-prompt across hundreds of models. Pick the right tool for the job (GPT, Claude, Gemini, image, video, code), pay only for what you use, walk away when you're done.
Two things make it different:
- Privacy. Prompts are encrypted in your browser before they leave you — not even PPQ sees them. No subscription tied to your identity, no log of your queries tied to your name.
- Optionality. One credential, every model. You don't get locked into one lab's roadmap or one provider's pricing. The model that's best for this job today might not be the one you're stuck paying $20/month for.
Together that's the foundation: a place to use AI that doesn't ask who you are or what you're working on. Jamie plugs into it.
Jamie: Plain English In, Real Audio Out
LLMs are great at reasoning. They're terrible about actually grounding things in reality.
They're terrible at "what did this specific person actually say on this specific show six months ago." Models will invent quotes with a straight face.
Jamie is built for exactly that gap, and it's two things stacked:
- Semantic search. Matches by meaning, not exact words. Ask in your own phrasing; Jamie finds the moments that match the idea, even if nobody on the show used your terms.
- NLP-tuned agent. You don't call search primitives or compose queries. You ask a question. Jamie plans the research — searching, filtering, pulling chapter context — and returns a synthesized answer with playable clips.
You don't get a paraphrase. You get the receipts — actual audio, actual timestamp, one click to play it back. Chatbots making up quotes is over. At least here.
For PPQ users that's a new dimension on the same counter: a tool you reach for the same way you reach for GPT or Claude, paid from the same PPQ balance, but grounded in what was actually said instead of what a model thinks was said. PPQ becomes the place to both think with a model and check what someone really said — one tab, one balance, one workflow.
L402: Hyperdeflation, Compounded
A line I keep coming back to:
> AI = hyperdeflation of implementing YOUR code.
> L402 = hyperdeflation of OTHER people's code.
LLMs let you write less of your own software. L402 lets you call other people's software per request, with a Lightning payment instead of a signup form, an API key, or an annual contract. No friction means you actually use the tool — and so does every agent you build.
PPQ and Jamie share that payment layer. The same balance that pays for a Claude prompt pays for a Jamie research call. Any agent you wire up — yours, or one of your customers' — can hit either one directly with a Lightning credential. No account, no key rotation, no dashboard. When you're done, you're done.
Try It
- Jamie on PPQ in your browser: ppq.ai
- Jamie's Agent Quickstart (LLMs.txt) directly via L402: pullthatupjamie.ai/llms.txt
- More on the agent stack: Pull That Up Jamie blog
FAQ
What is PPQ?
PPQ (PayPerQ) is a pay-per-prompt AI service — anonymous, no-subscription access to hundreds of AI models. Top up with Bitcoin, Lightning, crypto, or card; pay only for what you use.
What does Jamie add to PPQ?
Real spoken moments from podcasts. Hundreds of feeds, tens of thousands of episodes, millions of transcript paragraphs — searchable by meaning and orchestrated by an agent that takes your plain-English question and runs the research for you.
Why is Jamie better than asking ChatGPT the same question?
ChatGPT will invent quotes that sound right. Jamie only returns what was actually said, with audio you can click and verify.
Why is it better than searching Spotify or Apple Podcasts?
Those search titles and descriptions. Jamie searches the spoken words themselves — and matches on meaning, not exact keywords.
What does "semantic + NLP-tuned" mean in practice?
Two layers. Semantic search means meaning-based matching across transcripts. NLP-tuned means you don't compose queries — you ask a question and the agent figures out which searches to run, what to combine, and what to return.
Do I need an account?
No. PPQ doesn't require registration, and Jamie's API uses L402 Lightning auth.
What is L402?
Lightning-native HTTP auth. You pay once with Bitcoin Lightning, get a credential, and use it across endpoints until the balance runs out. No subscription, no email, no card on file.
Can I call Jamie from my own app or agent?
Yes. Hit /api/pull with an L402 credential — the same agent powering this on PPQ.
What's the "hyperdeflation" idea?
LLMs cut the cost of writing your own software. L402 cuts the cost of setup, coordination and paying for other people's software. Doing it in a discoverable pay per call, with no signup fashion. Stacked, the cost of building something useful with both drops fast.