How it works

Every claim carries its receipt.

Chart grades your brand as it is, writes the documents to fix it, and shows the source behind every claim it makes. This page is the mechanism: the same research engine behind our free research tools, packaged as the strategy deliverables agencies bill senior-partner rates for. Everything grounds in the same verified brand record as every Jinn product: how the record works.

Apply for early access

One rule governs this page: a research-backed claim cannot exist without a stored source it points to. That is not an editorial policy. It is a database constraint, described below.

At a glance
What you get

A catalog of 42 deliverable types across four depths: a grade on day one, then the operating documents, the researched landscape, and the senior synthesis read over your own completed reports.

How it is grounded

Every document reads from a Brand DNA record extracted from your own site and catalog, tracked field by field including what is absent, and generation is gated on that record being fresh within 30 days.

Why the receipts are real

22 of the 42 types run the research and verification spine. A claim in one literally cannot be stored without a source, because a database constraint enforces it, not application code that could be skipped.

What it refuses

To guess quietly. An unreadable verdict counts as unsupported, a report below the faithfulness gate does not publish, and in regulated categories an unverified sensitive claim is dropped rather than softened.

The receipts

Sourced, ledgered, verified, or dropped.

Four steps stand between a research-backed document and the reader, and each one is enforced in code or the database rather than trusted to a prompt.

  1. Sources storedEvery document the engine retrieves is saved before a single word of the report is written. The evidence exists first, so a claim can be checked against the thing it came from later.
  2. Claims ledgeredThe model states its facts as structured claims, each one citing a stored source. No citation, no claim: a claim row cannot be written with zero sources because a database constraint forbids it. Receipts are not a feature that can be turned off.
  3. Independently verifiedA second model reads each claim against the source text itself and returns a verdict. The report must clear a faithfulness gate of 0.85 supported claims. An unreadable or errored verdict is treated as unsupported, scored zero, never given the benefit of the doubt.
  4. Dropped, not softenedIn flagged regulated categories, a claim touching health, legal, financial, or age territory that is not fully source-verified is removed from the report entirely, and the suppression is written to an append-only audit log. Removed, never quietly reworded.

The other 20 deliverable types generate from your Brand DNA without the research appendix, and no pack page hides which is which. Receipts travel where the research does, and the page says so plainly.

The catalog

A grade today, the operating documents next.

The deliverables climb four depths, and a room never starts cold: the senior synthesis reads over your own completed reports rather than a blank page.

  • AuditThe grade. Your brand scored as it is today, including a coherence audit across 4 deterministic axes where every grade traces to the exact fields that produced it. Something real in hand on day one.
  • FoundationThe operating documents the grade says you are missing: the positioning, the messaging, the voice, written from your extracted record rather than a template.
  • IntelligenceThe landscape around you, mapped from competitors’ own published materials with the citations kept in the appendix rather than an analyst’s summary with the reading list left out.
  • Strategy RoomThe senior call, synthesized over everything the earlier tiers found. It unlocks after a completed earlier tier, so the recommendation is argued from your own reports, not a cold guess.
The machinery

The receipts are enforced, not promised.

The difference between Chart and a confident chatbot is that the honesty is structural. These are code and database facts, not editorial intentions.

A claim needs a source to existA research claim is stored with the sources it cites, and a database constraint rejects any claim that cites zero. The receipt is a precondition of the claim existing at all, enforced below the application where it cannot be bypassed.
The verifier fails closedA second model grades each claim against its source. If that verdict cannot be parsed, it is treated as unsupported and scored zero, and the report as a whole must clear a 0.85 faithfulness gate before it publishes. A broken judge lowers the score; it never waves a claim through.
The audit trail only growsRetrievals, verdicts, and regulated-claim suppressions are written to an append-only log that has no update or delete path at the API surface. The compliance record can grow, but it cannot be rewritten after the fact.
The appendix is the writer’s only source listThe numbered source appendix a reader sees is the same map the writing model was constrained to. The writer is forbidden from inventing a citation number that is not in the list, so the receipts a reader checks are the receipts the report was built from.
Apply for early access
Honest edges

A gap is labeled, not papered over.

Absent and present-but-low are different things, and the reports keep them different. When the engine did not have a signal, the report says so instead of rendering a confident number where a gap belongs. When the signal lands, the gap fills.

And the engine underneath is the same one behind our free research tools, not a separate Chart-only system. Chart is the packaging: the catalog, the tiers, and the reader. The grounding, the verification, and the provenance are the engine’s, and they behave the same wherever the engine runs.

Honest machinery

The receipts are a constraint, not a courtesy.

Receipts are a database ruleA research claim cannot be stored citing zero sources. The constraint lives in the database, below any code path, so the receipt is a precondition of the claim, not a step someone remembered to run.
Unverifiable means unsupportedAn unreadable verdict is scored as unsupported, never as passed. The conservative default is always the one that lowers the published score, so a parsing failure can never launder a claim into the report.
Regulated claims are droppedIn flagged regulated categories such as alcohol, cannabis, and supplements, an unverified sensitive claim is removed from the report entirely and the removal is logged. Dropped, never softened into something vaguer.
The record underneathEvery document grounds in the same verified brand record that powers every Jinn product: 346 fields structured from your own site and evidence, generation gated on that record being fresh. Chart reads the record. It does not invent one.
The substrate

Chart sits on the record.

The brand record every document grounds in is shared substrate, documented once and linked here rather than re-explained on every product page. The terms this page leans on are defined in the dictionary.

How the record worksFail-closed verificationProvenance
Questions

The questions people ask.

Which deliverables actually carry receipts?
22 of the 42 deliverable types run the research and verification spine today; the rest generate from your Brand DNA without the research appendix. No pack page hides which is which.
What does the source constraint actually mean?
It means a research claim is physically unable to be stored without citing at least one source. A database constraint rejects a claim with zero sources, so the receipts cannot be switched off or forgotten by the code.
What happens to health, legal, or financial claims?
For brands in flagged regulated categories, a claim in those sensitive areas that is not fully source-verified is dropped from the report entirely and the removal is written to an append-only audit log. Removed, never softened.
Is this a separate engine from your free tools?
No. Chart is the packaging around the same research engine that powers our free research tools: the catalog, the tiers, and the reader. The grounding and verification are the engine’s, and they behave the same wherever it runs.
How do I get access?
Chart is in private beta and purchasing is not open on the site. Request early access and we will bring you in as the catalog opens; prices publish with general access.

Strategy you can audit, claim by claim.

The grade, the roadmap, and the source behind every research-backed claim, with the receipts enforced in the database rather than promised in a footnote.

Apply for early access