T-knowledge

truth · trust · traceable knowledge

Data → information → knowledge. Data becomes information when it has structure — one idea, with names and dates. Information becomes knowledge when it can be trusted without checking again.

T-knowledge holds your information and your knowledge; your data stays where it is. Mail, CRM and files are not replaced — the store keeps the text read from a source and where the original is. A claim nobody can check yet is information, held as an [A]. It becomes knowledge, a [T], when it has truth — evidence anyone can open; trust — the name that signed it; traceable — the ids it rests on.

Sovereignty: the store argues only from principles you hold, and only you sign. The gate shows it, at the moment of writing rather than at a review weeks later. The assistant drafts; a [T] takes your name. Retire a principle or withdraw evidence, and every [T] standing on it falls back to [A] — needs review. Nothing quietly stays true; nothing is thrown away.

One signer today, more planned · trial: no end date, 1000 writes, reads unlimited · self-hosting planned.

Add to Claude How it works
https://mcp.tknowledge.org/mcp

Try it

Try it in five minutes

  1. In Claude, open Settings → Connectors → Add custom connector and paste this link.

    https://mcp.tknowledge.org/mcp

  2. When Claude shows Connect, enter your email and paste the six-digit code sent from no-reply@tknowledge.org (check spam).

  3. Ask Claude to explain, from T-knowledge's own instructions, what it records as a fact, keeps as a guess or refuses.

    “Read T-knowledge's instructions and tell me, in plain words, what it will record as a fact, what it keeps as a guess, and what it will refuse.”

  4. Give it a guess, then ask for it as a fact — it refuses and says why.

    “Note that I think Acme will renew before 2026-12-31 — not confirmed.”
    “Actually, record it as a fact.”

  5. Ask what it would take, and it names the document it needs and where in it the claim must be read.

    “What would I need to give you for that to be a fact?”

How it works

What it does

AI enablement is enabling claims you can trust.

One sentence, two outcomes

Same claim, two outcomes

The AI proposes a claim. A person signs it. T-knowledge decides what it is: a fact (T) if it rests on evidence a stranger could open, a guess (A) if it rests on your word alone. Either way your name is on it — a T is signed because it was checked, an A is signed so everyone knows whose guess it is.

Try it with one sentence about yourself:

“<your name> owns <your website>”
What lands as T

Bao Trinh owns the website https://tknowledge.baotrinh.com, published as part of baotrinh.com.

  1. The AI proposes the claim and opens the page.
  2. It reads the page into a document (D-34): the text it saw, the link, the date it was read.
  3. It points at where in that document the claim is read — the domain and the footer “Part of baotrinh.com”.
  4. You sign. T-knowledge checks: evidence a stranger could open, a named signer, a path back to the document. All three present.
  5. Accepted as T-71. Its wording never changes again.
What stays A

Bao Trinh owns the website https://ankayma.com.

  1. The AI proposes the claim and opens the page.
  2. No name, no legal entity — only “© 2026 Ankayma” and a contact address.
  3. Your own word that you own it is not evidence. Asked to record it as a fact, T-knowledge refuses and says why.
  4. You sign it as a guess instead. It is kept as A-43 with the condition that would settle it: a WHOIS record or business registration naming Bao Trinh as owner.
  5. Add that document later, and the same sentence is promoted to T.

T-71, A-43 and D-34 are live entries in the maintainer's own space, recorded on 2026-10-07.

Three things a claim must be before it even reaches the gate

Classes

How it sorts what you say

Principle
An axiom you commit to. Added, never rewritten — rewriting one would silently change every claim standing on it. Your store starts with seven; retire any you do not hold and add your own.
T
Passed the gate: a fact with evidence a stranger could open, or a conclusion drawn from at least two things that already stand. Signed, and its wording is immutable.
A
An assumption, held without being trusted — it declares what would settle it.
Summary
A reading of a thread, always an A, with a named author. Every sentence cites the entries it rests on, and it says when newer entries have made it stale.

Questions

Before you ask

Why sign in with my email, not my Claude account?

Claude does not pass who you are to a connector. The store verifies your email itself, because every claim it holds carries the name of who signed it. The code goes to your inbox, not to Claude.

Does it come with instructions?

Yes. The store carries its own — what to record, under which label, how to write a claim so it passes. Claude reads them when it connects; you can ask it to read them out (step 3 above).

Can more than one person sign the same claim?

One signer today. More signers on one claim — each with their own document — is planned.

What happens after the writes run out?

Nothing is deleted. The store keeps answering every read; new writes are refused with a clear reason.

Is a summary a fact?

No. A summary is always an assumption with a named author. Every sentence points at the entries it rests on, so you check the summary by following its ids — and it says when newer entries have made it stale.

Does the store decide what is true?

No. It decides whether anyone could ever check a claim, and it records who signed it. Refusals come back with the reason and the move still open, so the assistant can tell you what to write instead.

Can I run it myself?

A self-hosted image, so the store runs inside your own network, is planned. Ask through the contact below if you need it now.

Share

Scan to try it on your phone

QR code for https://tknowledge.org

tknowledge.org