educast.cc

Deep research, narrated.

A research pipeline that turns a topic into a long-form narrated audio episode — with chapters, notes, a cheat sheet, and a full source audit you can verify line-by-line. Built for the way you actually want to consume hard topics: while walking, driving, or doing the dishes.

What it is

educast.cc is the public face of an end-to-end research-to-audio pipeline. You hand it a subject and a few preferences — how dense, how long, how many modules, which language — and it does the work an analyst would do: maps the territory, gathers material, cross-checks the important claims against independent sources, plans a narrative, writes it, narrates it, and ships it as a self-contained shareable page.

Every share page carries the artifact and the receipts. You listen to the episode, read the notes, glance at the cheat sheet, and inspect the sources the writer was allowed to draw from — including which claims were independently corroborated and how strongly.

What you tell it

Four inputs shape an episode. Everything else is figured out by the pipeline.

Subject a prompt, optionally with required sub-topics Density how thorough you want the research Length roughly how long the episode should be Language any language you name in the prompt

How an episode comes together

A finished episode is the output of a sequence of specialised passes. Each pass is responsible for a different kind of judgement — finding things, checking things, structuring things, writing things, performing things. The pipeline is built so the steps stay honest: research feeds writing, writing feeds narration, and the source trail follows the whole way through.

1
Survey the subject
The pipeline reads your prompt, picks up any required sub-topics, and surveys the open web to figure out the shape of what is actually known. The output is a working understanding of what the topic contains and where the interesting parts live.
2
Go deep where it matters
For each part of the map, the pipeline goes back out and pulls additional material until it has enough substance to say something real. Your density and length preferences shape how thorough this gets.
3
Cross-check the important claims
Significant facts are independently re-verified against fresh sources before they are allowed near the script. Each claim ends up with a record: which sources support it, how many, and what an independent re-read concluded.
4
Plan the narrative
The verified material is shaped into a story arc rather than a wall of bullet points. Coverage you asked for is enforced; what was surfaced is allocated so no part of the subject gets starved.
5
Write the script
Long-form, sectioned, in the language you asked for. The writer follows a deliberate house style — direct, dry, light on the verbal tics that LLM prose usually drowns in — and is held to the points you required.
6
Narrate it like a human reads aloud
Text-to-speech with timing and prosody tuned for listening rather than skimming. Multiple voice providers are supported. Audio is produced chapter-by-chapter so playback can start before the full episode finishes rendering.
7
Package and publish
The audio, the notes, the cheat sheet, and the source audit are bundled into a single self-contained share page on educast.cc — player, waveform, chapter nav, offline cache, all in one place. No accounts, no tracking, no app.

Why you can trust what you hear

The cross-checking step is the part that earns the “deep research” in the name. A claim that surfaces once in one source is treated as provisional. A claim that survives an independent re-read — against a different set of sources than the ones that originally produced it — carries a higher confidence score and goes into the script with its corroboration count attached.

Every share page exposes this audit trail. For each fact you can see the original source, the number of independent sources that mention it, and whether the verification step re-checked it. This is not a guarantee of truth — nothing on the open web is — but it is a defensible epistemic floor. You can listen, then verify, then disagree, and the receipts are right there in the same page.

Languages

Name a language in your prompt — “in Hungarian”, “auf Deutsch”, “in Lithuanian”, anything similar — and the entire episode is produced in that language. There is no hardcoded language list to fall off the edge of; coverage is bounded only by what the underlying models and the chosen voice support, which in practice is a long list.

What you get with every episode

The audio

Narrated MP3 with a precise peaks waveform, variable speed, chapter markers, and offline caching for road trips and flights.

Notes

Structured notes the audio was written from — readable as a standalone briefing.

Cheat sheet

A compressed mental model: the hook, the entities, the spine of the argument, the anchor facts.

Sources & audit

Every source the research touched, with a credibility audit trail you can verify yourself.

What it’s good for

Got a topic you want researched? Pitch it.

Describe what you want an episode about and how thorough it should be. If a request is interesting it gets picked up and produced — you’ll get the share link by email if you provided one.

About the creator

educast.cc is built by Ben Racz — founder and managing director of a scientific research company, head of its quantitative analysis practice, and Professor of the Practice at Carnegie Mellon University. Career split across two-plus decades of complex software work: distributed research infrastructure, quantitative trading strategy validation, execution and order-management systems, risk and monitoring, and information security — ISO 27001 and Sarbanes-Oxley implementation, PCI-DSS technical consulting and vulnerability testing, and published research in cyberwarfare. Contributor to Humanity’s Last Exam, published in Nature. Reach out, give feedback, or follow the work at benracz.com.