Sources

Written by: Fatih

Every source we monitor is public and none of them is paid for. What we take is written here — and so is what we deliberately do not take, because the blind spots of an intelligence product matter as much as its coverage.

What we use

The full list of live channels, with what each one takes and does not take, is written out channel by channel in the evidence archive.

What we deliberately do not use

• Sources whose terms do not allow it. Some providers' free tiers are for “internal use only” or forbid commercial use outright. Those cannot be wired into a product, so they were not.
• Sources whose free path closed. During 2026 several notable sources removed free access or moved to pay-per-call. We do not keep them on a list hoping to “sort it out later”.
• Sources requiring written pre-approval. Not used without approval.

The limit on news feeds

From news feeds we take only the headline, link and date. Summaries and full text are neither stored nor republished; readers follow the link to the publisher's own page. This is both the lawful and the honest choice: we are not republishing the news, we are recording when something appeared.

What we do for machines — and what we do not

We would like to be cited as a source inside AI interfaces, but only through measured means. As far as we have measured: being indexable and quotable helps; keeping the source address in visible text helps (machine readers ignore embedded structured data during live retrieval); showing an author and a date helps. By contrast, adding a separate “AI file”, writing extra per-page schema, and simply publishing more pages have no measured effect — so we do none of them. The sources for these claims are recorded in our claim ledger as İ181–İ188.