The Silly Nice File.

The pitch for this piece, in our own editorial queue for most of a year, read: Silly Nice publishes full terpene panels and COAs on every SKU and is now carried in 150+ NY dispensaries. We map their catalog and place each cut. It was going to be the good-news volume — a New York craft brand doing voluntarily what our whole framework asks brands to do, and a chance to plot real terpene data in the Atlas for the first time in this series.

Two of the three premises did not survive reading the documents. The brand does not publish terpene panels on its product pages — it publishes terpene names, without percentages, on five of six SKUs. The dispensary count appears on the site as three different numbers simultaneously. And the framing that publishing terpene data is going beyond the requirement has it backwards: New York's Office of Cannabis Management mandates terpenoid testing and percentage-by-weight reporting on every adult-use batch. Every compliant New York brand has these panels. Silly Nice links to them, which is modestly above the median for consumer-facing brand sites in this market, and which is a different and smaller thing than inventing terpene transparency.

The third premise held, and it is the reason this volume is worth reading. The COAs are real, they come from three separately OCM-licensed laboratories, they carry batch numbers and Metrc package IDs, and the chemistry in them is genuinely good — better, in places, than the marketing built on top of it manages to convey. So we did map the catalog and place each cut. That analysis is the front half of this piece. The gap between the documents and the claims is the back half.

First, the part that works: the chemistry.

Six SKUs: a diamond-frosted infused flower, a THCa diamond powder, a bubble hash, a frosted hash ball, a 510 cartridge and a 2g all-in-one. Four have retrievable COAs with full terpene panels. Here is what those panels contain, and it is not what the product pages suggest.

Two numbers there deserve attention before anything else. 3.04% and 3.01% total terpenes on the two solventless products is a legitimately strong result. For context, the highest-terpene cultivar in the 2,257-strain Pestele dataset that powers our Atlas sits in the same neighborhood as well-made hash, and most commercial concentrate does not get there. Whoever is running the ice-water extraction is preserving aromatic mass properly.

The terpene-to-THC ratio, as a craft metric.

A number we find more useful than potency alone: how much aromatic mass a product carries per unit of THC. It separates products built for experience from products built for a number on a label, and it is computable from any compliant New York COA.

The two solventless products sit around 1:16–18, which is where good hash lives. The Diamond Powder sits at 1:862, which is not a criticism — a THCa isolate is supposed to be flavourless, that is the entire product concept — but it does mean the terpene story the brand tells across its range does not apply to its most potent SKU at all. All 21 terpene analytes on that COA read below the limit of quantitation.

And one thing the vape COA shows.

The 510 cartridge is marketed on 100% cannabis-derived terpenes. Its COA reports two detected terpenes: limonene at 1.072% and β-caryophyllene at 0.1061%. Limonene is 91% of the total terpene mass. Myrcene, α-pinene, β-pinene, humulene, linalool and terpinolene all read not-detected — while the product page advertises myrcene by name and says it "smooths the experience."

A full-spectrum cannabis-derived terpene fraction ordinarily retains eight to fifteen analytes above LOQ; the Bubble Hash on the same brand's shelf has seventeen. A two-analyte profile at 91:9 is the signature of a heavily refined or reconstructed blend. We are not in a position to say what went into that cart, and there are legitimate processes that strip a fraction down this far. What we can say is that the document does not support the myrcene claim printed above it.

Now the gap: claims against documents.

Everything above is drawn from the brand's own published COAs, which is possible only because the brand publishes them. That deserves acknowledgment before what follows. What follows is a set of places where the marketing layer has drifted from the paperwork underneath it — not, on this evidence, through any single deliberate act, but through the ordinary entropy of a small operation shipping fast and not going back to update pages when batches turn over.

Taken one at a time these are small. Taken together they describe a pattern worth naming, because it is common and because it is fixable: the marketing surface updates on a marketing cadence and the lab documents update on a batch cadence, and nobody owns reconciling the two. The consequence is that a brand genuinely doing the hard part — paying three labs, publishing the PDFs — ends up with pages that its own paperwork contradicts.

One item on that document set is not small, and it is not a marketing problem.

A reading on the Bubble Hash COA that we cannot reconcile.

The Bubble Hash certificate of analysis — ACT Laboratories NY, batch VHSNBH05, reported 7 May 2026, the same document that supplies the terpene profile praised earlier in this piece — contains a microbial panel whose reported values and whose summary line do not appear to agree.

On the face of the document, total aerobic bacteria is reported at roughly 23 times the limit printed beside it, and yeast and mold at roughly 2.3 times. Both rows are marked "Tested" rather than "Passed," while the panel's summary badge reads "Pass."

We are stating what the document shows and stopping there, deliberately. We are not asserting that anyone was sold unsafe product, that the laboratory erred, or that any party acted improperly. There are several explanations consistent with these figures that involve no wrongdoing at all: the total aerobic count may not be a pass/fail criterion for this product category under New York's testing rules, in which case reporting it without grading it is correct practice and the "Tested" label is doing exactly its job; the limits column may apply to a different product class than the one sampled; or the batch may have been remediated and retested under a document we did not locate. Any of those would resolve the apparent conflict entirely.

What we can say is that a reader who opens this COA to check the terpene numbers — which is precisely what the brand invites them to do, and what we did — will encounter two reported values above two printed limits under a heading that says Pass, with nothing on the document explaining why that is consistent. That is a legitimate thing for a customer to want explained, and the batch number is published here so anyone can look at the same page we did.

We flag it for one further reason. The entire argument of this publication is that a published document is worth more than a marketing claim, and that a reader should be able to check the chemistry themselves. A brand that publishes its COAs earns real credit for that, and Silly Nice does publish them. But publishing a document invites people to read all of it, not only the rows that flatter the product. That is the trade, and it is the right one.

A note on the information environment.

One structural fact is relevant to any reader trying to evaluate this brand from the public record, and we state it neutrally because it is a matter of filing, not of character.

A great deal of what appears in search results as independent coverage of Silly Nice originates from stupidDOPE, a lifestyle publication co-founded by one of the brand's own co-founders. STUPID DOPE LLC is registered with the New York Department of State (DOS ID 7371344, filed July 2024). By our count the outlet has published upwards of forty articles about the brand, alongside seventeen or more paid newswire releases.

Genuinely independent journalism about the brand amounts to roughly five items, none of them a standalone profile: an MJBizDaily policy story in which the brand is one of five operators quoted, a Leafly News roundup, a Cannabis Business Times roundup, a Ganjapreneur piece sourced to stupidDOPE, and a Harlem World Magazine item structured as a brand-supplied Q&A. We found no coverage in Green Market Report, Forbes, amNY, Crain's, High Times or Marijuana Moment.

There is nothing improper about founders operating a media property, and plenty of good brands are under-covered by trade press for reasons that have nothing to do with quality. The point is narrower and practical: a reader who encounters twelve enthusiastic articles about this brand may reasonably believe they have encountered twelve independent assessments. They have encountered one source with a consistent point of view. That is worth knowing before weighing it — and it is worth knowing here, too, since it is the reason our own original premise for this piece was wrong.

What this file is really about.

Seven volumes in, the failure modes have started to rhyme. BRĒZ published a mushroom dose in wet weight so a 33 mg extract could read as 2,200. FK+ said "no fairy dusting ever" over thirteen undisclosed botanicals. RIVR published every milligram and took a mixed review for it. Silly Nice publishes the documents and then lets the pages drift away from them.

Only one of those is a story about intent, and it isn't this one. The Silly Nice pattern is the most ordinary and probably the most widespread: a small operator doing the expensive, unglamorous part correctly — three labs, per-batch testing, public PDFs — and then losing the thread on the cheap part, which is keeping the words next to the numbers true as the numbers change.

The reason a published standard matters is that it makes that drift visible without anyone having to allege anything. We did not need to know what the company meant. We only needed the COA and the page, side by side. Every claim in the table above was checkable in an afternoon by anyone with a browser, which is exactly the property a market needs more of.