The price watch

When a vendor moves a price, this page says so

Every figure in our price index was read off a vendor's own page on one day and then frozen. Vendors do not freeze. So each morning a job re-reads the 43 vendor pricing pages behind the index — 42 tools, 68 published figures — and diffs what they say today against what we quoted. Anything that no longer matches is recorded below, dated, with the excerpt it was read from — and anything else on the web still citing the old figure goes into the correction queue beneath it.

3 open corrections · 0 third-party pages named · the queue

We diff the vendor's OWN published figures, the ones our briefs quote directly — never our annualised arithmetic, which no vendor page contains. The job reads HTML and declared JSON-LD, not JavaScript: where a rate card renders client-side and declares nothing, the run is recorded as "could not read its prices" rather than as a change. A withdrawal — "this figure is gone" — additionally requires the page to publish real rates in the same unit today: where it publishes none, we cannot tell a withdrawn price from a page we could not read, and we say the second. "We could not read it" and "they changed it" are different claims and this page keeps them apart. And where a vendor publishes no machine-readable rate in the unit we quote at all, that is not an unread page either: it is declared below, with every candidate page we fetched and the date, rather than left as a daily blank. Last completed pass: 2026-10-25. Runs every morning at 05:40 UTC.

3 open changes

Vendor price moved · first seen 2026-10-23 · confirmed on 11 runs

Intercom moved the Essential seat figure our Fin brief quotes down from $39 to $35 (-10.3%).

Read off the vendor's own published page, in the same unit our brief quotes it in. Our annual figure for this tool is computed from the old rate and has not been rewritten yet.

What we published
$39 · index row: $30,000–$45,000/year
What the page says today
$35 · -10.3%
Read from
https://www.intercom.com/pricing · HTTP 200 · last checked 2026-10-25
The sentence to correct
helpdesk seats at $39 / $99 / $139 per seat per month

…lot is an optional add-on at $29 per agent/month billed annually , or $35 per agent/month billed monthly, for unlimited usage. Copilot is a personal AI a…

Our Fin write-up, and the working behind the figure →

Figure no longer published · first seen 2026-10-20 · confirmed on 2 runs

canny.io's own pricing page no longer publishes the $660 Third-party (per month) figure our Canny Autopilot brief quotes, and publishes no comparable figure in its place.

A withdrawn figure means the number our brief quotes is no longer on the vendor's page — which can mean a rise, a cut, or a move behind a quote or a calculator. It is a flag to re-research, not a new price.

What we published
$660 · index row: $7,900/year
What the page says today
no comparable figure
Read from
https://canny.io/pricing · HTTP 200 · last checked 2026-10-20
The sentence to correct
Third-party reporting puts a 20-person org near $660 a month at roughly 1,000 tracked users, which is about $7,900 a year — our arithmetic (×12) on a reported monthly figure, not a Canny number.

…Nearest figure on the page today: $79 — tracked users Automatic feedback capture Unlimited feedback starts at $ 79 / month billed yearly Pro Connect feedback to revenue with advanced reporting a…

Our Canny Autopilot write-up, and the working behind the figure →

Vendor price moved · first seen 2026-10-18 · confirmed on 20 runs

figma.com moved the top-ups (per month) figure our Figma Agent brief quotes down from $150 to $90 (-40%).

Read off the vendor's own published page, in the same unit our brief quotes it in. Our annual figure for this tool is computed from the old rate and has not been rewritten yet.

What we published
$150 · index row: $4,800/year
What the page says today
$90 · -40%
Read from
https://www.figma.com/pricing · HTTP 200 · last checked 2026-10-25
The sentence to correct
Organization plan at $55 per Full seat per month, annual only, because SAML SSO starts there; a 20-person product org typically runs about 6 Full and 14 Collab seats ($5) = roughly $4,800/year, plus AI credit top-ups published at $150/month per 5,000 shared credits, so $5,000–$6,600/year.

…more information? Contact sales Enterprise Billed annually Full seat $90 /mo + 4,250 AI credits/mo Dev seat $35 /mo + 500 AI credits/mo Collab seat $5 /…

Our Figma Agent write-up, and the working behind the figure →

3 records we withdrew

A monitor that quietly deletes its own mistakes is not auditable. When a record here turns out to be wrong, it comes off the list above and is named here instead, with the date and what we got wrong — and no correction is sent off the back of it.

WithdrawnThe claim, and why it was wrong
2026-10-18[email protected] · event-rate · figure_withdrawn
False positive. The $0.0000343 product-analytics event rate is still published, unchanged: https://posthog.com/product-analytics/pricing, 'Calculate your cost' table, '2 million-15 million $0.0000343/event' (HTTP 200, read by hand 2026-10-18). The watched page, posthog.com/pricing, renders its rates in a client-side calculator and declared only $0 placeholders; the watch has been repointed and the unreadable-page guard tightened so an empty calculator can no longer be read as a withdrawn price.
2026-10-18[email protected] · per-month-750 · figure_withdrawn
False positive, same cause: the Scale package is still published at $750/month on https://posthog.com/platform-packages (HTTP 200, read 2026-10-18). posthog.com/pricing paints that figure client-side and never served it to us.
2026-10-18[email protected] · per-month-250 · figure_withdrawn
False positive, same cause: the Boost package is still published at $250/month on https://posthog.com/platform-packages (HTTP 200, read 2026-10-18).

10 rows we corrected by hand

The watch above can only see figures a vendor's page still publishes. It cannot see a figure that was never the vendor's in the first place — our own ×12 of a monthly rate, or a different product's plan price sitting in the same paragraph. Those are found by reading, and every one we find is rewritten against the vendor's live page and listed here: what the table printed before, what it prints now, the page that was read and the day it was read. The figure changes on every surface at once — the table, the CSV, the dataset descriptors and the tool page — because they all read the same corrected paragraph.

A row can appear twice. 11 rounds are listed below across 10 rows, because a figure can be wrong in more than one way and each pass is published separately rather than overwritten: Sentry Seer was corrected on 19 October for quoting the wrong line item, and again on 20 October because the rate was right but the unit was not — $40 is charged per active contributor (any user opening 2+ pull requests a month to a Seer-enabled repository, counted fresh each month), which is not a seat, so multiplying it by 20 seats stated a precision we do not have. Where we cannot state a figure in the unit we claim, the row says so and carries a band or a named ceiling instead of a number that looks firmer than it is. The site serves the newest round; the older ones stay here as the record.

CorrectedToolWas → now, and why
2026-10-20Sentry Seer$9,600/year, tagged our arithmetic — $40 × 20 seats × 12 → $2,880–$9,600/year, our arithmetic, with the contributor count named
The $40 rate was right and the MULTIPLIER was not. Sentry's billing documentation, re-read today, defines an active contributor as any user who makes two or more pull requests in a month to a Seer-enabled repository, with the count resetting monthly — it is not a seat. Multiplying it by 20 silently asserted that every person in a 20-person PRODUCT org (PMs, designers, EMs included) opens 2+ PRs a month into a Seer-enabled repo, which is close to certainly false. The row is now a band whose floor assumes six active contributors and whose ceiling assumes all twenty, with the ceiling named as a ceiling, and the paragraph states the 2-PR definition in full.
Read https://docs.sentry.io/pricing/ on 2026-10-20: “Access Seer by adding it to your existing subscription for $40 per active contributor per month, where an active contributor is defined as any user who makes 2 or more PRs to a Seer-Enabled repository” (docs.sentry.io/pricing, Seer Pricing). sentry.io/pricing now serves the same $40/active contributor/month in HTML, and docs.sentry.io/product/ai-in-sentry/seer/ repeats it as “any person who creates 2 or more PRs/MRs in a month in a Seer-Enabled repo/project will be billed”.
2026-10-19Productboard Spark$7,080/year, labelled buyer-reported → $12,000–$24,000/year, buyer-reported
$7,080 was our own arithmetic — $59 per maker per month × 10 makers × 12 — printed as though buyers had reported it, and the $59 Pro seat rate it rests on is no longer published anywhere on Productboard's pricing page. The brief's own conclusion was already that a 20-person org lands on Enterprise, which is custom-quoted, so the row now carries the reseller-and-buyer band for that tier and says whose figure it is.
Read https://www.productboard.com/pricing on 2026-10-19: Business (“Request demo”, 25 contributors, full Spark access) and Enterprise (“Contact sales”). No seat rate at all; the only figures our reader finds on the page are credit top-ups — $5 per 50-credit bundle monthly, $60 per 600-credit bundle annually.
2026-10-19Hex Threads$9,000/year, off a brief quoting “$900/editor/year” as a Hex figure → $9,000/year, with both steps of the arithmetic named as ours
Hex publishes a monthly seat rate and no annual one. The $900 per editor per year the brief quoted is $75 × 12 — our multiplication, written as though Hex printed it — and the $9,000 row is that again × 10 editors. The figure is unchanged because the monthly rate still holds (verified on the page today); what changes is that nothing here is now attributed to Hex that Hex does not publish.
Read https://hex.tech/pricing on 2026-10-19: Professional $36 per editor per month, Team $75 per editor per month, Enterprise custom. No annual figure anywhere on the page.
2026-10-19Sentry Seer$960/year — our ×12 of the Sentry Business plan, on a row about Seer → $9,600/year — our ×12 of Seer's own published per-contributor rate
$960 is Sentry Business at $80 a month annualised: the PLAN Seer is added to, not Seer. Sentry publishes Seer at $40 per active contributor per month, so twenty active contributors is $9,600 a year by our own arithmetic. The independent write-ups quoting $40 per active contributor are right and our row was quoting a different line item — note that sentry.io/pricing paints the Seer block client-side, so the figure is only readable in HTML on sentry.io/product/seer/ and in Sentry's pricing docs.
Read https://sentry.io/product/seer/ on 2026-10-19: “Seer is available to all Sentry users on a paid plan (Team, Business, or Enterprise) at $40 per active contributor, per month” (sentry.io/product/seer/, server-rendered; repeated in docs.sentry.io/pricing). Team $26/mo, Business $80/mo billed annually.
2026-10-19Productboard Spark$60/year, labelled vendor-published → $8,000–$20,000/year, buyer-reported
$60 is the price of a 600-credit top-up bundle — a unit rate the extractor read as this org's annual cost and then labelled as Productboard's own published yearly price. The paragraph's real answer, the band comparable buyers report, was sitting in the next sentence. The unit-rate guard in lib/tool-price.ts now covers “per 600 credits” as well, so a counted unit cannot be read as an org total again.
Read https://www.productboard.com/pricing on 2026-10-19: Business (“Request demo”) and Enterprise (“Contact sales”); the only figures on the page are credit top-ups at $5 per 50-credit bundle monthly and $60 per 600-credit bundle annually.
2026-10-19Pylon$16,700/year, labelled buyer-reported, off “$89 per seat annually” → $16,700/year, our arithmetic on Pylon's own published seat rates
Two errors in one sentence: the $16,700 is our own multiplication, not a figure any tracker published, and the rates behind it are not annual — $89 and $50 a seat only reach $16,700 across ten seats if they are MONTHLY. Pylon publishes seat rates again today (Professional $89/seat, Advanced $139/seat, which is the $89 + $50 the brief described), with no billing period printed beside them, so the monthly reading is ours and is now labelled as ours.
Read https://www.usepylon.com/pricing on 2026-10-19: Professional $89/seat and Advanced $139/seat with no period printed; add-ons on the same page are monthly — SAML SSO $125/mo, Phone $35 per seat/month, Account Intelligence $10 per customer account.
2026-10-19Zendesk AI Agent Conversation Simulator$13,800/year, labelled buyer-reported → $13,800/year, our arithmetic on Zendesk's published seat rate
Nobody reported $13,800 to us: it is Zendesk's published $115 per agent per month × 10 seats × 12, where the ten shared seats are our assumption too. The rate itself holds on the page today; the label was the error.
Read https://www.zendesk.com/pricing on 2026-10-19: Suite Professional $115 per agent/month on annual billing ($149 monthly), Suite Growth $55, Suite Team $19; Copilot add-on $50 per agent/month.
2026-10-19Canny Autopilot$7,900/year, labelled buyer-reported → $7,900/year, our ×12 of a reported monthly figure
The $660 a month at 1,000 tracked users is third-party reporting — Canny's own slider renders client-side and the page serves only the $79 entry price — but the $7,900 is our ×12 of it, which the row attributed to the reporters as well.
Read https://canny.io/pricing on 2026-10-19: Free $0, Pro “starts at $79/month billed yearly”, Business “Let's talk”. The tracked-user slider that prices Pro above $79 is client-side and serves no figure.
2026-10-19QA.tech$6,000/year, labelled buyer-reported → $6,000/year, our ×12 of a reported monthly rate
qa.tech/pricing publishes no figure at all — our own reader finds zero dollar amounts on it — so the $499 a month is a third-party comparison site's number and the $6,000 is our twelve-month multiplication of it. Only the first half of that was labelled.
Read https://qa.tech/pricing on 2026-10-19: Every tier, Starter included, is quote-only behind a demo; the page publishes no dollar figure in HTML or JSON-LD.
2026-10-19GoodData.AI Agentic Serving Plane$29,940/year, labelled buyer-reported → $29,940/year, our ×12 of a third-party listing
$29,940 is $2,495 × 12. The monthly figure comes from comparable published listings — gooddata.ai/pricing itself publishes no figure our reader can find — but the annualisation is ours, and the row credited the whole number to the listings.
Read https://www.gooddata.ai/pricing on 2026-10-19: No dollar figure in HTML or JSON-LD on gooddata.ai/pricing, and no line item for the Agentic Serving Plane.
2026-10-19Atlassian Rovo$4,450/year, our arithmetic on seat rates Atlassian has since moved → $4,780/year, the same arithmetic on the rates published today
Flagged on 18 Oct as arithmetic a diff could not safely rewrite. Re-read by hand today: Jira Premium is $15.56 per user per month (was $14.54) and Atlassian Guard Standard $4.35 (was about $4.00–$4.20), both from Atlassian's own pricing pages, which puts 20 seats at $398.20 a month — about $4,780 a year. The two open watch records for those moves are correct and close themselves once the brief quotes the new rates.
Read https://www.atlassian.com/software/jira/pricing on 2026-10-19: Jira Standard $8.48 and Premium $15.56 per user per month; Atlassian Guard Standard $4.35 and Premium $8.55 per user per month; Rovo credit overage $0.01 per credit ($10.00 per 1,000 credits).

3 open corrections

A moved price does not only make OUR row stale. The same figure is usually sitting in other people's pricing round-ups, and the morning the vendor moves it, every one of those pages is wrong too. So each change above opens an entry here naming the pages on record citing the old figure, the sentence each states it in, and the route we checked for reaching that publisher. Nothing is sent automatically: this is a dated work list with the vendor's own page attached as the evidence.

Observed live on 2026-10-17 by searching for each figure and reading the page back: 14 third-party pages across 5 figures. The citation rounds recorded earlier this month live in growth documents this build could not read, so nothing here is copied from them — every entry carries the URL, the sentence and the date it was read.

No third-party page on record · opened 2026-10-23 · seen on 11 runs

Intercom moved the Essential seat figure our Fin brief quotes down from $39 to $35 (-10.3%). No third-party page is on record citing our $39 figure.

The figure
$39 · Essential seat
What the vendor publishes now
$35 · -10.3%
Read from
https://www.intercom.com/pricing · last checked 2026-10-25
Our own pages carrying it
/tools/fin · /ai-tool-prices · /ai-tool-prices.csv

Nothing in our citation ledger quotes this figure, so this entry is our own correction work only: the brief, the index row and the CSV above.

The note, with the evidence in it

Correction, with the source: Intercom's own pricing page (https://www.intercom.com/pricing) now publishes $35 where it published $39. We read it on 2026-10-25 (HTTP 200); the excerpt is "lot is an optional add-on at $29 per agent/month billed annually , or $35 per agent/month billed monthly, for unlimited usage. Copilot is a personal AI a". We published $39 too, and our own record of the move — old figure, new figure, date, excerpt — is at https://signal-brief19.acoco.ai/price-changes#c-b4921566. No link or credit needed; we would just rather the number were right. — Signal Brief ([email protected])

No third-party page on record · opened 2026-10-20 · seen on 2 runs

canny.io's own pricing page no longer publishes the $660 Third-party (per month) figure our Canny Autopilot brief quotes, and publishes no comparable figure in its place. No third-party page is on record citing our $660 figure.

The figure
$660 · Third-party (per month)
What the vendor publishes now
no comparable figure
Read from
https://canny.io/pricing · last checked 2026-10-20

Nothing in our citation ledger quotes this figure, so this entry is our own correction work only: the brief, the index row and the CSV above.

The note, with the evidence in it

Correction, with the source: canny.io's own pricing page (https://canny.io/pricing) no longer publishes the $660 Third-party (per month) figure. We read it on 2026-10-20 (HTTP 200); the excerpt is "Nearest figure on the page today: $79 — tracked users Automatic feedback capture Unlimited feedback starts at $ 79 / month billed yearly Pro Connect feedback to". We published $660 too, and our own record of the move — old figure, new figure, date, excerpt — is at https://signal-brief19.acoco.ai/price-changes#c-fade6d73. No link or credit needed; we would just rather the number were right. — Signal Brief ([email protected])

No third-party page on record · opened 2026-10-18 · seen on 20 runs

figma.com moved the top-ups (per month) figure our Figma Agent brief quotes down from $150 to $90 (-40%). No third-party page is on record citing our $150 figure.

The figure
$150 · top-ups (per month)
What the vendor publishes now
$90 · -40%
Read from
https://www.figma.com/pricing · last checked 2026-10-25

Nothing in our citation ledger quotes this figure, so this entry is our own correction work only: the brief, the index row and the CSV above.

The note, with the evidence in it

Correction, with the source: figma.com's own pricing page (https://www.figma.com/pricing) now publishes $90 where it published $150. We read it on 2026-10-25 (HTTP 200); the excerpt is "more information? Contact sales Enterprise Billed annually Full seat $90 /mo + 4,250 AI credits/mo Dev seat $35 /mo + 500 AI credits/mo Collab seat $5 /". We published $150 too, and our own record of the move — old figure, new figure, date, excerpt — is at https://signal-brief19.acoco.ai/price-changes#c-ae6f3154. No link or credit needed; we would just rather the number were right. — Signal Brief ([email protected])

5 figures no rate card publishes

"Could not read its prices" is a statement about US, and on 19 October it was covering for three different things: a reader bug on our side, a page we were pointed at wrongly, and a vendor who simply does not publish the number our brief quotes. The first two are fixed in code. The third is not a failure to read anything — it is a fact about the vendor — so each one is declared here instead: the figure, every candidate rate card we fetched, what each one publishes instead, and the date it was read. These figures are no longer counted as unread pages, and they are not diffed, because there is nothing on any page to diff them against.

2 of them — Lorikeet, Sprig Analyze Agent — look like figures that may have gone stale rather than figures a vendor never published, so they are queued for a human to re-research. A declaration is not a licence to stop looking.

ReadToolThe figure, and why there is nothing to diff
2026-10-20Lorikeet
needs a hand re-read
$0.10 per ticket (Auto QA) — no rate card we can find publishes it
Lorikeet prices per resolved action, not per ticket. Its Quality SKU — the Auto QA line our brief quotes at “from $0.10 per ticket” — is published at $0.30 on Start and $0.27 on Scale, per action, on both the pricing page and the machine-readable card. No per-ticket figure exists on either, so there is nothing for the watch to compare our $0.10 against.
Read https://www.lorikeetcx.ai/pricing (HTTP 200) on 2026-10-20: plan prices Start $2,100/mo and Scale $5,100/mo (both still match our figures) and per-action rates — Triage $0.30 / $0.27, Resolution $0.99 / $0.90, Voice $1.50 / $1.20, Automated QA $0.30 / $0.27 — none of them quoted per ticket.
Read https://www.lorikeetcx.ai/pricing.md (HTTP 200) on 2026-10-20: Lorikeet's own machine-readable rate card, linked from the pricing page as “Pricing also available in machine-readable format”: the same three SKUs (Triage, Resolution, Quality) priced per action, with no $0.10 and no per-ticket unit anywhere in it.
The two Lorikeet figures that ARE on the card ($2,100/mo and $5,100/mo) stay watched and both hold. The Auto QA line needs a human to re-read and rewrite the brief sentence against the published per-action rates; until then this row says so rather than reporting the whole page unread.
2026-10-20Hex Threads
nothing to re-check
$900 per editor, per year — no rate card we can find publishes it
Hex publishes a monthly seat rate and no annual one. The $900 per editor per year this row used to quote is our own ×12 of the $75 monthly rate, which our brief already says in as many words — so the watch was looking for a figure Hex has never published.
Read https://hex.tech/pricing (HTTP 200) on 2026-10-20: monthly seat rates only — Team $75 per Editor per month and Professional $36 per Editor per month, in the page text and again in declared JSON-LD — plus per-minute compute rates; Enterprise is custom-quoted and no annual seat figure appears anywhere.
The two rates Hex does publish ($75 and $36 per editor per month) are watched and both hold. Our annual headline stays as our stated arithmetic on the monthly rate, which is what the index row already says.
2026-10-20Productboard Spark
nothing to re-check
$59 per maker, per month — no rate card we can find publishes it
Productboard withdrew its published seat pricing. The $59 per maker per month Pro rate this row was built on is not on the rate card in any form, and the only prices it publishes are credit top-up bundles — which are not a year of the product and not what our row quotes.
Read https://www.productboard.com/pricing (HTTP 200) on 2026-10-20: two tiers and no seat rate — Business behind “Request demo”, Enterprise behind “Contact sales”, a 5-maker minimum stated in words — and, in declared JSON-LD, credit top-up bundles only: $5 per 50 credits monthly and $60 per 600 credits annually.
Read https://www.productboard.com/pricing/ (HTTP 200) on 2026-10-20: the same page, same two tiers, same credit bundles, no seat rate.
Already corrected by hand on 19 Oct: the index row is now a buyer-reported band, labelled as buyer-reported, and the brief says the seat rate is gone. With no vendor-published figure left, the row has no watch target — named here rather than counted as a page we could not read.
2026-10-20Sprig Analyze Agent
needs a hand re-read
$175 per month (Starter, billed annually) — no rate card we can find publishes it
Sprig publishes no price anywhere on its own site today. Its pricing page carries no dollar figure in its HTML or its JSON-LD, and its documentation states plan pricing varies by need, so there is no machine-readable rate for the watch to read, let alone diff.
Read https://sprig.com/pricing (HTTP 200) on 2026-10-20: no figure at all — 3,561 characters of server-rendered text, every plan behind “Book a demo”, and the page now describes a single enterprise platform whose “pricing scales with your research program”; the Starter tier is not named on it.
Read https://sprig.com/plans (HTTP 200) on 2026-10-20: the same page, byte for byte.
Read https://sprig.com/pricing/ (HTTP 200) on 2026-10-20: the same page, byte for byte.
Read https://docs.sprig.com/docs/account-and-settings/plans-teams-and-billing (HTTP 200) on 2026-10-20: the Starter plan's quotas (400 replay clips, 200 in-product survey responses, up to 50k MUUs) and the sentence “Plan pricing varies based on needs” — no price.
Our $175/month Starter figure can no longer be re-verified on a Sprig page, and the tier it names is no longer listed. Flagged for a hand re-read; the index row's own headline is the buyer-reported Enterprise band, which was never a Sprig list price and is labelled as such.
2026-10-20Lovable Cloud
nothing to re-check
$50 per month (Business, per workspace) — no rate card we can find publishes it
Lovable's plan prices exist only after JavaScript runs. Every page on their site that we can read publishes credit rates, never the $50 a month Business price our row is built on, so no URL we could point the watch at would make that figure diffable.
Read https://lovable.dev/pricing (HTTP 200) on 2026-10-20: no figure at all: the plan cards are painted client-side, and the string “$50” does not appear anywhere in the 184 KB of HTML served to us — not in the markup, not in a script payload, not in JSON-LD. The server-rendered text is the FAQ and the word “Plans”.
Read https://docs.lovable.dev/introduction/plans-and-credits (HTTP 200) on 2026-10-20: credit conversion rates only — Pro $0.25 per credit, Business $0.50 per credit — and a link back to lovable.dev/pricing for plan prices.
The figure stays in the index as read on the day it was researched, and this declaration stands in for a daily check nobody can perform. If Lovable ever server-renders its plan table, the row goes back under the watch.

What is watched

One row per vendor page: 13 typed in by hand against the brief they came from, 30 derived automatically from the researched rows of the index, so a tool priced in this morning's issue is watched tomorrow morning. The figures column is what our own brief quotes from that page — the only numbers the watch compares, and the test suite fails if any of them has drifted from the sentence in the brief it was taken from.

ToolVendor pageFigures watchedLast read
Momentichttps://momentic.ai/pricing$125, $0.01875matches our figure
2026-10-25
Finhttps://fin.ai/pricing$0.99, $9.99could not read its prices
2026-10-25
Finhttps://www.intercom.com/pricing$39, $99, $139does not match
2026-10-25
PostHog AIhttps://posthog.com/product-analytics/pricing$0.0000343matches our figure
2026-10-25
Atlassian Rovohttps://www.atlassian.com/software/jira/pricing$8.48, $15.56matches our figure
2026-10-25
Listen Labshttps://listenlabs.ai/articles/ai-qualitative-research-pricing/$25, $50matches our figure
2026-10-25
Linear for Agentshttps://linear.app/pricing$16, $10matches our figure
2026-10-25
Notion Agentshttps://www.notion.com/pricing$20, $10matches our figure
2026-10-25
Mintlify Assistanthttps://www.mintlify.com/pricing$450, $0.01matches our figure
2026-10-25
GrowthBook 5.0 Product Analyticshttps://www.growthbook.io/pricing$40, $30matches our figure
2026-10-25
LaunchDarklyhttps://launchdarkly.com/pricing/$10, $8.33matches our figure
2026-10-25
Votitohttps://www.votito.com/pricing$29matches our figure
2026-10-25
Dovetail — Digital Twins and Agentshttps://dovetail.com/pricing/no rate card published (we watch for one appearing)matches our figure
2026-10-25
Alloyhttps://alloy.app/pricing$20matches our figure
2026-10-25
Braintrusthttps://www.braintrust.dev/pricing$249matches our figure
2026-10-25
Bug0https://bug0.com/pricing$2,500matches our figure
2026-10-25
Canny Autopilothttps://canny.io/pricing$79matches our figure
2026-10-25
ChatPRDhttps://www.chatprd.ai/pricing$29, $349, $15matches our figure
2026-10-25
Confident AI Customer Monitoringhttps://www.confident-ai.com/pricing$200, $2,000matches our figure
2026-10-25
cubichttps://www.cubic.dev/pricing$40matches our figure
2026-10-25
Decagon Duet Apprenticehttps://decagon.ai/pricing$0.50, $0.99did not answer
2026-10-25
Figma Agenthttps://www.figma.com/pricing$150does not match
2026-10-25
Frankhttps://hifrank.ai/pricing$312matches our figure
2026-10-25
Front Compasshttps://front.com/pricing$65, $0.05, $20, $105matches our figure
2026-10-25
Graphite Agenthttps://graphite.com/pricing$40matches our figure
2026-10-25
Hex Notebook Agenthttps://hex.tech/pricing$75matches our figure
2026-10-25
Hex Threadshttps://hex.tech/pricing$75, $36matches our figure
2026-10-25
Intercom Fin 3https://fin.ai/pricing$0.99matches our figure
2026-10-25
LangSmith Insights Agenthttps://www.langchain.com/pricing$39matches our figure
2026-10-25
Linear Product Intelligencehttps://linear.app/pricing$16matches our figure
2026-10-25
Lorikeethttps://www.lorikeetcx.ai/pricing$2,100, $5,100matches our figure
2026-10-25
Machahttps://www.getmacha.com/pricing$1,199matches our figure
2026-10-25
Maze AI Moderatorhttps://maze.co/pricing$99could not read its prices
2026-10-25
New Relic AI Evaluationhttps://newrelic.com/pricing$349, $49matches our figure
2026-10-25
Perspective AIhttps://getperspective.ai/pricing$99matches our figure
2026-10-25
PostHog AIhttps://posthog.com/platform-packages$750, $250matches our figure
2026-10-25
Pylonhttps://www.usepylon.com/pricing$125, $35matches our figure
2026-10-25
Replit Agent 3https://replit.com/pricing$20matches our figure
2026-10-25
Sentry Seerhttps://sentry.io/pricing$26, $80matches our figure
2026-10-25
TestCollab AI Agentshttps://testcollab.com/pricing$39matches our figure
2026-10-25
Unified.to Agent Skillshttps://unified.to/pricing$750, $3,000matches our figure
2026-10-25
Vercel Agenthttps://vercel.com/pricing$20, $300matches our figure
2026-10-25
Zendesk AI Agent Conversation Simulatorhttps://www.zendesk.com/pricing$115matches our figure
2026-10-25

What is not watched, and why

Not every researched row has a vendor figure to diff. Where our price comes from buyer reports or our own arithmetic on units the vendor states in prose, there is no published number on a rate card for a monitor to compare — and where no conventional rate-card URL answered at all, we say that rather than invent a page to watch. Both are named here instead of being quietly left out of the count.

ToolWhy it is not watched
Amplitude AI Agentsno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Conveono vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Dovetail AI Agentsno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Dovetail Digital Twinsno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Enterpret Agent OSno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Fullstory StoryAI Agentsno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
GoodData.AI Agentic Serving Planeno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Inkeep Agentsno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Kojino vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
LaunchDarkly AgentControlno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
LaunchDarkly AgentControlno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Lovable Cloudthe rate card publishes no such figure (declared above)
Lovable's plan prices exist only after JavaScript runs. Every page on their site that we can read publishes credit rates, never the $50 a month Business price our row is built on, so no URL we could point the watch at would make that figure diffable. Pages read on 2026-10-20: https://lovable.dev/pricing, https://docs.lovable.dev/introduction/plans-and-credits.
Mixpanel AIno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
OpenAI AgentKitno rate card we could find
no candidate path answered 200
Pendo for Agentsno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Pendo Novusno rate card we could find
no candidate path answered 200
Productboard Sparkthe rate card publishes no such figure (declared above)
Productboard withdrew its published seat pricing. The $59 per maker per month Pro rate this row was built on is not on the rate card in any form, and the only prices it publishes are credit top-up bundles — which are not a year of the product and not what our row quotes. Pages read on 2026-10-20: https://www.productboard.com/pricing, https://www.productboard.com/pricing/.
Productboard Sparkno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
QA Wolf Mapping AIno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
QA.techno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Rangerno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Sprig Analyze Agentthe rate card publishes no such figure (declared above)
Sprig publishes no price anywhere on its own site today. Its pricing page carries no dollar figure in its HTML or its JSON-LD, and its documentation states plan pricing varies by need, so there is no machine-readable rate for the watch to read, let alone diff. Pages read on 2026-10-20: https://sprig.com/pricing, https://sprig.com/plans, https://sprig.com/pricing/, https://docs.sprig.com/docs/account-and-settings/plans-teams-and-billing.
Spurno vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page
Strellano vendor-published figure in our brief
the stored brief quotes no vendor-published rate for this tool — the figure we publish is our arithmetic or a buyer-reported band, and neither is on a vendor page

How a change becomes a correction

A record here is evidence, not a rewrite. The brief behind it stays exactly as published until a human re-reads the vendor's page and rewrites it, because the paragraph carries caveats and arithmetic a diff cannot safely edit. What the record does is make the correction uncontestable: the vendor's page, the date, the old figure, the new one. Spotted something this missed? [email protected] — a dated vendor URL is all we need.

The table those figures live in is the price index, and the same rows as data are in the CSV. Both carry the figures as published; this page is where you find out which of them the vendor has since contradicted.

7 weekday issues free · no card · nothing renews

Prices move. Get the five tools, priced, every weekday — 7 issues free, no card.

The five tools are published free every weekday morning. Subscribers also get the Adoption Brief on each one: adopt / pilot / skip, the price a 20-person product org really lands on, the SSO and data-risk line, and a pitch you can paste into #product. Leave your email and the next 7 issues arrive with all five briefs attached — no card, nothing to cancel.

  • A one-word verdict on every tool — adopt, pilot or skip — with the reason in one sentence.
  • The real price a 20-person product org lands on, not the headline tier on the pricing page.
  • The integration and data-risk line: which plan SSO needs, the SOC 2 posture, where your roadmap and customer data ends up.
  • A paste-ready 3-line pitch for #product, so making the case costs you nothing.

Read a real Adoption Brief →

Free for 7 issues
no card · no password · nothing renews

Rather skip the trial and subscribe now? $10/month, or $96/year paid once.

No card to start the trial · 7 weekday issues with every brief · then the five tools stay free · $10/month or $96/year keeps the briefs · stop any time by emailing [email protected]

Signal Brief · the only thing we sell

Tomorrow morning, five more — with the call on each.

Every issue is published here in full, free, and stays that way. What a subscription adds is the Adoption Brief on all five tools: adopt, pilot or skip and why, the price a 20-person product org really lands on, the SSO and data-risk line, and a 3-line pitch you can paste into #product.

7 weekday issues free · no card · nothing renews
  • A one-word verdict on every tool — adopt, pilot or skip — with the reason in one sentence.
  • The real price a 20-person product org lands on, not the headline tier on the pricing page.
  • The integration and data-risk line: which plan SSO needs, the SOC 2 posture, where your roadmap and customer data ends up.
  • A paste-ready 3-line pitch for #product, so making the case costs you nothing.

Read a real Adoption Brief →

Free for 7 issues
no card · no password · nothing renews

Rather skip the trial and subscribe now? $10/month, or $96/year paid once.

No card to start the trial · 7 weekday issues with every brief · then the five tools stay free · $10/month or $96/year keeps the briefs · stop any time by emailing [email protected]

$10/month or $96/year after 7 free days, cancel in one click · a real Adoption Brief, in full · every issue, published free · [email protected]